cleanup: prefer matchmod.{always,never}() over accessing matchers directly

We have these factory methods so we should use them.

Differential Revision: https://phab.mercurial-scm.org/D5927

mercurial/changegroup.py

@@ -1,1418 +1,1418 b''
 # changegroup.py - Mercurial changegroup manipulation functions
 #
 #  Copyright 2006 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import os
 import struct
 import weakref
 
 from .i18n import _
 from .node import (
     hex,
     nullid,
     nullrev,
     short,
 )
 
 from . import (
     error,
     match as matchmod,
     mdiff,
     phases,
     pycompat,
     repository,
     util,
 )
 
 _CHANGEGROUPV1_DELTA_HEADER = struct.Struct("20s20s20s20s")
 _CHANGEGROUPV2_DELTA_HEADER = struct.Struct("20s20s20s20s20s")
 _CHANGEGROUPV3_DELTA_HEADER = struct.Struct(">20s20s20s20s20sH")
 
 LFS_REQUIREMENT = 'lfs'
 
 readexactly = util.readexactly
 
 def getchunk(stream):
     """return the next chunk from stream as a string"""
     d = readexactly(stream, 4)
     l = struct.unpack(">l", d)[0]
     if l <= 4:
         if l:
             raise error.Abort(_("invalid chunk length %d") % l)
         return ""
     return readexactly(stream, l - 4)
 
 def chunkheader(length):
     """return a changegroup chunk header (string)"""
     return struct.pack(">l", length + 4)
 
 def closechunk():
     """return a changegroup chunk header (string) for a zero-length chunk"""
     return struct.pack(">l", 0)
 
 def _fileheader(path):
     """Obtain a changegroup chunk header for a named path."""
     return chunkheader(len(path)) + path
 
 def writechunks(ui, chunks, filename, vfs=None):
     """Write chunks to a file and return its filename.
 
     The stream is assumed to be a bundle file.
     Existing files will not be overwritten.
     If no filename is specified, a temporary file is created.
     """
     fh = None
     cleanup = None
     try:
         if filename:
             if vfs:
                 fh = vfs.open(filename, "wb")
             else:
                 # Increase default buffer size because default is usually
                 # small (4k is common on Linux).
                 fh = open(filename, "wb", 131072)
         else:
             fd, filename = pycompat.mkstemp(prefix="hg-bundle-", suffix=".hg")
             fh = os.fdopen(fd, r"wb")
         cleanup = filename
         for c in chunks:
             fh.write(c)
         cleanup = None
         return filename
     finally:
         if fh is not None:
             fh.close()
         if cleanup is not None:
             if filename and vfs:
                 vfs.unlink(cleanup)
             else:
                 os.unlink(cleanup)
 
 class cg1unpacker(object):
     """Unpacker for cg1 changegroup streams.
 
     A changegroup unpacker handles the framing of the revision data in
     the wire format. Most consumers will want to use the apply()
     method to add the changes from the changegroup to a repository.
 
     If you're forwarding a changegroup unmodified to another consumer,
     use getchunks(), which returns an iterator of changegroup
     chunks. This is mostly useful for cases where you need to know the
     data stream has ended by observing the end of the changegroup.
 
     deltachunk() is useful only if you're applying delta data. Most
     consumers should prefer apply() instead.
 
     A few other public methods exist. Those are used only for
     bundlerepo and some debug commands - their use is discouraged.
     """
     deltaheader = _CHANGEGROUPV1_DELTA_HEADER
     deltaheadersize = deltaheader.size
     version = '01'
     _grouplistcount = 1 # One list of files after the manifests
 
     def __init__(self, fh, alg, extras=None):
         if alg is None:
             alg = 'UN'
         if alg not in util.compengines.supportedbundletypes:
             raise error.Abort(_('unknown stream compression type: %s')
                              % alg)
         if alg == 'BZ':
             alg = '_truncatedBZ'
 
         compengine = util.compengines.forbundletype(alg)
         self._stream = compengine.decompressorreader(fh)
         self._type = alg
         self.extras = extras or {}
         self.callback = None
 
     # These methods (compressed, read, seek, tell) all appear to only
     # be used by bundlerepo, but it's a little hard to tell.
     def compressed(self):
         return self._type is not None and self._type != 'UN'
     def read(self, l):
         return self._stream.read(l)
     def seek(self, pos):
         return self._stream.seek(pos)
     def tell(self):
         return self._stream.tell()
     def close(self):
         return self._stream.close()
 
     def _chunklength(self):
         d = readexactly(self._stream, 4)
         l = struct.unpack(">l", d)[0]
         if l <= 4:
             if l:
                 raise error.Abort(_("invalid chunk length %d") % l)
             return 0
         if self.callback:
             self.callback()
         return l - 4
 
     def changelogheader(self):
         """v10 does not have a changelog header chunk"""
         return {}
 
     def manifestheader(self):
         """v10 does not have a manifest header chunk"""
         return {}
 
     def filelogheader(self):
         """return the header of the filelogs chunk, v10 only has the filename"""
         l = self._chunklength()
         if not l:
             return {}
         fname = readexactly(self._stream, l)
         return {'filename': fname}
 
     def _deltaheader(self, headertuple, prevnode):
         node, p1, p2, cs = headertuple
         if prevnode is None:
             deltabase = p1
         else:
             deltabase = prevnode
         flags = 0
         return node, p1, p2, deltabase, cs, flags
 
     def deltachunk(self, prevnode):
         l = self._chunklength()
         if not l:
             return {}
         headerdata = readexactly(self._stream, self.deltaheadersize)
         header = self.deltaheader.unpack(headerdata)
         delta = readexactly(self._stream, l - self.deltaheadersize)
         node, p1, p2, deltabase, cs, flags = self._deltaheader(header, prevnode)
         return (node, p1, p2, cs, deltabase, delta, flags)
 
     def getchunks(self):
         """returns all the chunks contains in the bundle
 
         Used when you need to forward the binary stream to a file or another
         network API. To do so, it parse the changegroup data, otherwise it will
         block in case of sshrepo because it don't know the end of the stream.
         """
         # For changegroup 1 and 2, we expect 3 parts: changelog, manifestlog,
         # and a list of filelogs. For changegroup 3, we expect 4 parts:
         # changelog, manifestlog, a list of tree manifestlogs, and a list of
         # filelogs.
         #
         # Changelog and manifestlog parts are terminated with empty chunks. The
         # tree and file parts are a list of entry sections. Each entry section
         # is a series of chunks terminating in an empty chunk. The list of these
         # entry sections is terminated in yet another empty chunk, so we know
         # we've reached the end of the tree/file list when we reach an empty
         # chunk that was proceeded by no non-empty chunks.
 
         parts = 0
         while parts < 2 + self._grouplistcount:
             noentries = True
             while True:
                 chunk = getchunk(self)
                 if not chunk:
                     # The first two empty chunks represent the end of the
                     # changelog and the manifestlog portions. The remaining
                     # empty chunks represent either A) the end of individual
                     # tree or file entries in the file list, or B) the end of
                     # the entire list. It's the end of the entire list if there
                     # were no entries (i.e. noentries is True).
                     if parts < 2:
                         parts += 1
                     elif noentries:
                         parts += 1
                     break
                 noentries = False
                 yield chunkheader(len(chunk))
                 pos = 0
                 while pos < len(chunk):
                     next = pos + 2**20
                     yield chunk[pos:next]
                     pos = next
             yield closechunk()
 
     def _unpackmanifests(self, repo, revmap, trp, prog):
         self.callback = prog.increment
         # no need to check for empty manifest group here:
         # if the result of the merge of 1 and 2 is the same in 3 and 4,
         # no new manifest will be created and the manifest group will
         # be empty during the pull
         self.manifestheader()
         deltas = self.deltaiter()
         repo.manifestlog.getstorage(b'').addgroup(deltas, revmap, trp)
         prog.complete()
         self.callback = None
 
     def apply(self, repo, tr, srctype, url, targetphase=phases.draft,
               expectedtotal=None):
         """Add the changegroup returned by source.read() to this repo.
         srctype is a string like 'push', 'pull', or 'unbundle'.  url is
         the URL of the repo where this changegroup is coming from.
 
         Return an integer summarizing the change to this repo:
         - nothing changed or no source: 0
         - more heads than before: 1+added heads (2..n)
         - fewer heads than before: -1-removed heads (-2..-n)
         - number of heads stays the same: 1
         """
         repo = repo.unfiltered()
         def csmap(x):
             repo.ui.debug("add changeset %s\n" % short(x))
             return len(cl)
 
         def revmap(x):
             return cl.rev(x)
 
         changesets = files = revisions = 0
 
         try:
             # The transaction may already carry source information. In this
             # case we use the top level data. We overwrite the argument
             # because we need to use the top level value (if they exist)
             # in this function.
             srctype = tr.hookargs.setdefault('source', srctype)
             tr.hookargs.setdefault('url', url)
             repo.hook('prechangegroup',
                       throw=True, **pycompat.strkwargs(tr.hookargs))
 
             # write changelog data to temp files so concurrent readers
             # will not see an inconsistent view
             cl = repo.changelog
             cl.delayupdate(tr)
             oldheads = set(cl.heads())
 
             trp = weakref.proxy(tr)
             # pull off the changeset group
             repo.ui.status(_("adding changesets\n"))
             clstart = len(cl)
             progress = repo.ui.makeprogress(_('changesets'), unit=_('chunks'),
                                             total=expectedtotal)
             self.callback = progress.increment
 
             efiles = set()
             def onchangelog(cl, node):
                 efiles.update(cl.readfiles(node))
 
             self.changelogheader()
             deltas = self.deltaiter()
             cgnodes = cl.addgroup(deltas, csmap, trp, addrevisioncb=onchangelog)
             efiles = len(efiles)
 
             if not cgnodes:
                 repo.ui.develwarn('applied empty changelog from changegroup',
                                   config='warn-empty-changegroup')
             clend = len(cl)
             changesets = clend - clstart
             progress.complete()
             self.callback = None
 
             # pull off the manifest group
             repo.ui.status(_("adding manifests\n"))
             # We know that we'll never have more manifests than we had
             # changesets.
             progress = repo.ui.makeprogress(_('manifests'), unit=_('chunks'),
                                             total=changesets)
             self._unpackmanifests(repo, revmap, trp, progress)
 
             needfiles = {}
             if repo.ui.configbool('server', 'validate'):
                 cl = repo.changelog
                 ml = repo.manifestlog
                 # validate incoming csets have their manifests
                 for cset in pycompat.xrange(clstart, clend):
                     mfnode = cl.changelogrevision(cset).manifest
                     mfest = ml[mfnode].readdelta()
                     # store file cgnodes we must see
                     for f, n in mfest.iteritems():
                         needfiles.setdefault(f, set()).add(n)
 
             # process the files
             repo.ui.status(_("adding file changes\n"))
             newrevs, newfiles = _addchangegroupfiles(
                 repo, self, revmap, trp, efiles, needfiles)
             revisions += newrevs
             files += newfiles
 
             deltaheads = 0
             if oldheads:
                 heads = cl.heads()
                 deltaheads = len(heads) - len(oldheads)
                 for h in heads:
                     if h not in oldheads and repo[h].closesbranch():
                         deltaheads -= 1
             htext = ""
             if deltaheads:
                 htext = _(" (%+d heads)") % deltaheads
 
             repo.ui.status(_("added %d changesets"
                              " with %d changes to %d files%s\n")
                              % (changesets, revisions, files, htext))
             repo.invalidatevolatilesets()
 
             if changesets > 0:
                 if 'node' not in tr.hookargs:
                     tr.hookargs['node'] = hex(cl.node(clstart))
                     tr.hookargs['node_last'] = hex(cl.node(clend - 1))
                     hookargs = dict(tr.hookargs)
                 else:
                     hookargs = dict(tr.hookargs)
                     hookargs['node'] = hex(cl.node(clstart))
                     hookargs['node_last'] = hex(cl.node(clend - 1))
                 repo.hook('pretxnchangegroup',
                           throw=True, **pycompat.strkwargs(hookargs))
 
             added = [cl.node(r) for r in pycompat.xrange(clstart, clend)]
             phaseall = None
             if srctype in ('push', 'serve'):
                 # Old servers can not push the boundary themselves.
                 # New servers won't push the boundary if changeset already
                 # exists locally as secret
                 #
                 # We should not use added here but the list of all change in
                 # the bundle
                 if repo.publishing():
                     targetphase = phaseall = phases.public
                 else:
                     # closer target phase computation
 
                     # Those changesets have been pushed from the
                     # outside, their phases are going to be pushed
                     # alongside. Therefor `targetphase` is
                     # ignored.
                     targetphase = phaseall = phases.draft
             if added:
                 phases.registernew(repo, tr, targetphase, added)
             if phaseall is not None:
                 phases.advanceboundary(repo, tr, phaseall, cgnodes)
 
             if changesets > 0:
 
                 def runhooks():
                     # These hooks run when the lock releases, not when the
                     # transaction closes. So it's possible for the changelog
                     # to have changed since we last saw it.
                     if clstart >= len(repo):
                         return
 
                     repo.hook("changegroup", **pycompat.strkwargs(hookargs))
 
                     for n in added:
                         args = hookargs.copy()
                         args['node'] = hex(n)
                         del args['node_last']
                         repo.hook("incoming", **pycompat.strkwargs(args))
 
                     newheads = [h for h in repo.heads()
                                 if h not in oldheads]
                     repo.ui.log("incoming",
                                 "%d incoming changes - new heads: %s\n",
                                 len(added),
                                 ', '.join([hex(c[:6]) for c in newheads]))
 
                 tr.addpostclose('changegroup-runhooks-%020i' % clstart,
                                 lambda tr: repo._afterlock(runhooks))
         finally:
             repo.ui.flush()
         # never return 0 here:
         if deltaheads < 0:
             ret = deltaheads - 1
         else:
             ret = deltaheads + 1
         return ret
 
     def deltaiter(self):
         """
         returns an iterator of the deltas in this changegroup
 
         Useful for passing to the underlying storage system to be stored.
         """
         chain = None
         for chunkdata in iter(lambda: self.deltachunk(chain), {}):
             # Chunkdata: (node, p1, p2, cs, deltabase, delta, flags)
             yield chunkdata
             chain = chunkdata[0]
 
 class cg2unpacker(cg1unpacker):
     """Unpacker for cg2 streams.
 
     cg2 streams add support for generaldelta, so the delta header
     format is slightly different. All other features about the data
     remain the same.
     """
     deltaheader = _CHANGEGROUPV2_DELTA_HEADER
     deltaheadersize = deltaheader.size
     version = '02'
 
     def _deltaheader(self, headertuple, prevnode):
         node, p1, p2, deltabase, cs = headertuple
         flags = 0
         return node, p1, p2, deltabase, cs, flags
 
 class cg3unpacker(cg2unpacker):
     """Unpacker for cg3 streams.
 
     cg3 streams add support for exchanging treemanifests and revlog
     flags. It adds the revlog flags to the delta header and an empty chunk
     separating manifests and files.
     """
     deltaheader = _CHANGEGROUPV3_DELTA_HEADER
     deltaheadersize = deltaheader.size
     version = '03'
     _grouplistcount = 2 # One list of manifests and one list of files
 
     def _deltaheader(self, headertuple, prevnode):
         node, p1, p2, deltabase, cs, flags = headertuple
         return node, p1, p2, deltabase, cs, flags
 
     def _unpackmanifests(self, repo, revmap, trp, prog):
         super(cg3unpacker, self)._unpackmanifests(repo, revmap, trp, prog)
         for chunkdata in iter(self.filelogheader, {}):
             # If we get here, there are directory manifests in the changegroup
             d = chunkdata["filename"]
             repo.ui.debug("adding %s revisions\n" % d)
             deltas = self.deltaiter()
             if not repo.manifestlog.getstorage(d).addgroup(deltas, revmap, trp):
                 raise error.Abort(_("received dir revlog group is empty"))
 
 class headerlessfixup(object):
     def __init__(self, fh, h):
         self._h = h
         self._fh = fh
     def read(self, n):
         if self._h:
             d, self._h = self._h[:n], self._h[n:]
             if len(d) < n:
                 d += readexactly(self._fh, n - len(d))
             return d
         return readexactly(self._fh, n)
 
 def _revisiondeltatochunks(delta, headerfn):
     """Serialize a revisiondelta to changegroup chunks."""
 
     # The captured revision delta may be encoded as a delta against
     # a base revision or as a full revision. The changegroup format
     # requires that everything on the wire be deltas. So for full
     # revisions, we need to invent a header that says to rewrite
     # data.
 
     if delta.delta is not None:
         prefix, data = b'', delta.delta
     elif delta.basenode == nullid:
         data = delta.revision
         prefix = mdiff.trivialdiffheader(len(data))
     else:
         data = delta.revision
         prefix = mdiff.replacediffheader(delta.baserevisionsize,
                                          len(data))
 
     meta = headerfn(delta)
 
     yield chunkheader(len(meta) + len(prefix) + len(data))
     yield meta
     if prefix:
         yield prefix
     yield data
 
 def _sortnodesellipsis(store, nodes, cl, lookup):
     """Sort nodes for changegroup generation."""
     # Ellipses serving mode.
     #
     # In a perfect world, we'd generate better ellipsis-ified graphs
     # for non-changelog revlogs. In practice, we haven't started doing
     # that yet, so the resulting DAGs for the manifestlog and filelogs
     # are actually full of bogus parentage on all the ellipsis
     # nodes. This has the side effect that, while the contents are
     # correct, the individual DAGs might be completely out of whack in
     # a case like 882681bc3166 and its ancestors (back about 10
     # revisions or so) in the main hg repo.
     #
     # The one invariant we *know* holds is that the new (potentially
     # bogus) DAG shape will be valid if we order the nodes in the
     # order that they're introduced in dramatis personae by the
     # changelog, so what we do is we sort the non-changelog histories
     # by the order in which they are used by the changelog.
     key = lambda n: cl.rev(lookup(n))
     return sorted(nodes, key=key)
 
 def _resolvenarrowrevisioninfo(cl, store, ischangelog, rev, linkrev,
                                linknode, clrevtolocalrev, fullclnodes,
                                precomputedellipsis):
     linkparents = precomputedellipsis[linkrev]
     def local(clrev):
         """Turn a changelog revnum into a local revnum.
 
         The ellipsis dag is stored as revnums on the changelog,
         but when we're producing ellipsis entries for
         non-changelog revlogs, we need to turn those numbers into
         something local. This does that for us, and during the
         changelog sending phase will also expand the stored
         mappings as needed.
         """
         if clrev == nullrev:
             return nullrev
 
         if ischangelog:
             return clrev
 
         # Walk the ellipsis-ized changelog breadth-first looking for a
         # change that has been linked from the current revlog.
         #
         # For a flat manifest revlog only a single step should be necessary
         # as all relevant changelog entries are relevant to the flat
         # manifest.
         #
         # For a filelog or tree manifest dirlog however not every changelog
         # entry will have been relevant, so we need to skip some changelog
         # nodes even after ellipsis-izing.
         walk = [clrev]
         while walk:
             p = walk[0]
             walk = walk[1:]
             if p in clrevtolocalrev:
                 return clrevtolocalrev[p]
             elif p in fullclnodes:
                 walk.extend([pp for pp in cl.parentrevs(p)
                                 if pp != nullrev])
             elif p in precomputedellipsis:
                 walk.extend([pp for pp in precomputedellipsis[p]
                                 if pp != nullrev])
             else:
                 # In this case, we've got an ellipsis with parents
                 # outside the current bundle (likely an
                 # incremental pull). We "know" that we can use the
                 # value of this same revlog at whatever revision
                 # is pointed to by linknode. "Know" is in scare
                 # quotes because I haven't done enough examination
                 # of edge cases to convince myself this is really
                 # a fact - it works for all the (admittedly
                 # thorough) cases in our testsuite, but I would be
                 # somewhat unsurprised to find a case in the wild
                 # where this breaks down a bit. That said, I don't
                 # know if it would hurt anything.
                 for i in pycompat.xrange(rev, 0, -1):
                     if store.linkrev(i) == clrev:
                         return i
                 # We failed to resolve a parent for this node, so
                 # we crash the changegroup construction.
                 raise error.Abort(
                     'unable to resolve parent while packing %r %r'
                     ' for changeset %r' % (store.indexfile, rev, clrev))
 
         return nullrev
 
     if not linkparents or (
         store.parentrevs(rev) == (nullrev, nullrev)):
         p1, p2 = nullrev, nullrev
     elif len(linkparents) == 1:
         p1, = sorted(local(p) for p in linkparents)
         p2 = nullrev
     else:
         p1, p2 = sorted(local(p) for p in linkparents)
 
     p1node, p2node = store.node(p1), store.node(p2)
 
     return p1node, p2node, linknode
 
 def deltagroup(repo, store, nodes, ischangelog, lookup, forcedeltaparentprev,
                topic=None,
                ellipses=False, clrevtolocalrev=None, fullclnodes=None,
                precomputedellipsis=None):
     """Calculate deltas for a set of revisions.
 
     Is a generator of ``revisiondelta`` instances.
 
     If topic is not None, progress detail will be generated using this
     topic name (e.g. changesets, manifests, etc).
     """
     if not nodes:
         return
 
     cl = repo.changelog
 
     if ischangelog:
         # `hg log` shows changesets in storage order. To preserve order
         # across clones, send out changesets in storage order.
         nodesorder = 'storage'
     elif ellipses:
         nodes = _sortnodesellipsis(store, nodes, cl, lookup)
         nodesorder = 'nodes'
     else:
         nodesorder = None
 
     # Perform ellipses filtering and revision massaging. We do this before
     # emitrevisions() because a) filtering out revisions creates less work
     # for emitrevisions() b) dropping revisions would break emitrevisions()'s
     # assumptions about delta choices and we would possibly send a delta
     # referencing a missing base revision.
     #
     # Also, calling lookup() has side-effects with regards to populating
     # data structures. If we don't call lookup() for each node or if we call
     # lookup() after the first pass through each node, things can break -
     # possibly intermittently depending on the python hash seed! For that
     # reason, we store a mapping of all linknodes during the initial node
     # pass rather than use lookup() on the output side.
     if ellipses:
         filtered = []
         adjustedparents = {}
         linknodes = {}
 
         for node in nodes:
             rev = store.rev(node)
             linknode = lookup(node)
             linkrev = cl.rev(linknode)
             clrevtolocalrev[linkrev] = rev
 
             # If linknode is in fullclnodes, it means the corresponding
             # changeset was a full changeset and is being sent unaltered.
             if linknode in fullclnodes:
                 linknodes[node] = linknode
 
             # If the corresponding changeset wasn't in the set computed
             # as relevant to us, it should be dropped outright.
             elif linkrev not in precomputedellipsis:
                 continue
 
             else:
                 # We could probably do this later and avoid the dict
                 # holding state. But it likely doesn't matter.
                 p1node, p2node, linknode = _resolvenarrowrevisioninfo(
                     cl, store, ischangelog, rev, linkrev, linknode,
                     clrevtolocalrev, fullclnodes, precomputedellipsis)
 
                 adjustedparents[node] = (p1node, p2node)
                 linknodes[node] = linknode
 
             filtered.append(node)
 
         nodes = filtered
 
     # We expect the first pass to be fast, so we only engage the progress
     # meter for constructing the revision deltas.
     progress = None
     if topic is not None:
         progress = repo.ui.makeprogress(topic, unit=_('chunks'),
                                         total=len(nodes))
 
     configtarget = repo.ui.config('devel', 'bundle.delta')
     if configtarget not in ('', 'p1', 'full'):
         msg = _("""config "devel.bundle.delta" as unknown value: %s""")
         repo.ui.warn(msg % configtarget)
 
     deltamode = repository.CG_DELTAMODE_STD
     if forcedeltaparentprev:
         deltamode = repository.CG_DELTAMODE_PREV
     elif configtarget == 'p1':
         deltamode = repository.CG_DELTAMODE_P1
     elif configtarget == 'full':
         deltamode = repository.CG_DELTAMODE_FULL
 
     revisions = store.emitrevisions(
         nodes,
         nodesorder=nodesorder,
         revisiondata=True,
         assumehaveparentrevisions=not ellipses,
         deltamode=deltamode)
 
     for i, revision in enumerate(revisions):
         if progress:
             progress.update(i + 1)
 
         if ellipses:
             linknode = linknodes[revision.node]
 
             if revision.node in adjustedparents:
                 p1node, p2node = adjustedparents[revision.node]
                 revision.p1node = p1node
                 revision.p2node = p2node
                 revision.flags |= repository.REVISION_FLAG_ELLIPSIS
 
         else:
             linknode = lookup(revision.node)
 
         revision.linknode = linknode
         yield revision
 
     if progress:
         progress.complete()
 
 class cgpacker(object):
     def __init__(self, repo, oldmatcher, matcher, version,
                  builddeltaheader, manifestsend,
                  forcedeltaparentprev=False,
                  bundlecaps=None, ellipses=False,
                  shallow=False, ellipsisroots=None, fullnodes=None):
         """Given a source repo, construct a bundler.
 
         oldmatcher is a matcher that matches on files the client already has.
         These will not be included in the changegroup.
 
         matcher is a matcher that matches on files to include in the
         changegroup. Used to facilitate sparse changegroups.
 
         forcedeltaparentprev indicates whether delta parents must be against
         the previous revision in a delta group. This should only be used for
         compatibility with changegroup version 1.
 
         builddeltaheader is a callable that constructs the header for a group
         delta.
 
         manifestsend is a chunk to send after manifests have been fully emitted.
 
         ellipses indicates whether ellipsis serving mode is enabled.
 
         bundlecaps is optional and can be used to specify the set of
         capabilities which can be used to build the bundle. While bundlecaps is
         unused in core Mercurial, extensions rely on this feature to communicate
         capabilities to customize the changegroup packer.
 
         shallow indicates whether shallow data might be sent. The packer may
         need to pack file contents not introduced by the changes being packed.
 
         fullnodes is the set of changelog nodes which should not be ellipsis
         nodes. We store this rather than the set of nodes that should be
         ellipsis because for very large histories we expect this to be
         significantly smaller.
         """
         assert oldmatcher
         assert matcher
         self._oldmatcher = oldmatcher
         self._matcher = matcher
 
         self.version = version
         self._forcedeltaparentprev = forcedeltaparentprev
         self._builddeltaheader = builddeltaheader
         self._manifestsend = manifestsend
         self._ellipses = ellipses
 
         # Set of capabilities we can use to build the bundle.
         if bundlecaps is None:
             bundlecaps = set()
         self._bundlecaps = bundlecaps
         self._isshallow = shallow
         self._fullclnodes = fullnodes
 
         # Maps ellipsis revs to their roots at the changelog level.
         self._precomputedellipsis = ellipsisroots
 
         self._repo = repo
 
         if self._repo.ui.verbose and not self._repo.ui.debugflag:
             self._verbosenote = self._repo.ui.note
         else:
             self._verbosenote = lambda s: None
 
     def generate(self, commonrevs, clnodes, fastpathlinkrev, source,
                  changelog=True):
         """Yield a sequence of changegroup byte chunks.
         If changelog is False, changelog data won't be added to changegroup
         """
 
         repo = self._repo
         cl = repo.changelog
 
         self._verbosenote(_('uncompressed size of bundle content:\n'))
         size = 0
 
         clstate, deltas = self._generatechangelog(cl, clnodes,
                                                   generate=changelog)
         for delta in deltas:
             for chunk in _revisiondeltatochunks(delta,
                                                 self._builddeltaheader):
                 size += len(chunk)
                 yield chunk
 
         close = closechunk()
         size += len(close)
         yield closechunk()
 
         self._verbosenote(_('%8.i (changelog)\n') % size)
 
         clrevorder = clstate['clrevorder']
         manifests = clstate['manifests']
         changedfiles = clstate['changedfiles']
 
         # We need to make sure that the linkrev in the changegroup refers to
         # the first changeset that introduced the manifest or file revision.
         # The fastpath is usually safer than the slowpath, because the filelogs
         # are walked in revlog order.
         #
         # When taking the slowpath when the manifest revlog uses generaldelta,
         # the manifest may be walked in the "wrong" order. Without 'clrevorder',
         # we would get an incorrect linkrev (see fix in cc0ff93d0c0c).
         #
         # When taking the fastpath, we are only vulnerable to reordering
         # of the changelog itself. The changelog never uses generaldelta and is
         # never reordered. To handle this case, we simply take the slowpath,
         # which already has the 'clrevorder' logic. This was also fixed in
         # cc0ff93d0c0c.
 
         # Treemanifests don't work correctly with fastpathlinkrev
         # either, because we don't discover which directory nodes to
         # send along with files. This could probably be fixed.
         fastpathlinkrev = fastpathlinkrev and (
             'treemanifest' not in repo.requirements)
 
         fnodes = {}  # needed file nodes
 
         size = 0
         it = self.generatemanifests(
             commonrevs, clrevorder, fastpathlinkrev, manifests, fnodes, source,
             clstate['clrevtomanifestrev'])
 
         for tree, deltas in it:
             if tree:
                 assert self.version == b'03'
                 chunk = _fileheader(tree)
                 size += len(chunk)
                 yield chunk
 
             for delta in deltas:
                 chunks = _revisiondeltatochunks(delta, self._builddeltaheader)
                 for chunk in chunks:
                     size += len(chunk)
                     yield chunk
 
             close = closechunk()
             size += len(close)
             yield close
 
         self._verbosenote(_('%8.i (manifests)\n') % size)
         yield self._manifestsend
 
         mfdicts = None
         if self._ellipses and self._isshallow:
             mfdicts = [(self._repo.manifestlog[n].read(), lr)
                        for (n, lr) in manifests.iteritems()]
 
         manifests.clear()
         clrevs = set(cl.rev(x) for x in clnodes)
 
         it = self.generatefiles(changedfiles, commonrevs,
                                 source, mfdicts, fastpathlinkrev,
                                 fnodes, clrevs)
 
         for path, deltas in it:
             h = _fileheader(path)
             size = len(h)
             yield h
 
             for delta in deltas:
                 chunks = _revisiondeltatochunks(delta, self._builddeltaheader)
                 for chunk in chunks:
                     size += len(chunk)
                     yield chunk
 
             close = closechunk()
             size += len(close)
             yield close
 
             self._verbosenote(_('%8.i  %s\n') % (size, path))
 
         yield closechunk()
 
         if clnodes:
             repo.hook('outgoing', node=hex(clnodes[0]), source=source)
 
     def _generatechangelog(self, cl, nodes, generate=True):
         """Generate data for changelog chunks.
 
         Returns a 2-tuple of a dict containing state and an iterable of
         byte chunks. The state will not be fully populated until the
         chunk stream has been fully consumed.
 
         if generate is False, the state will be fully populated and no chunk
         stream will be yielded
         """
         clrevorder = {}
         manifests = {}
         mfl = self._repo.manifestlog
         changedfiles = set()
         clrevtomanifestrev = {}
 
         state = {
             'clrevorder': clrevorder,
             'manifests': manifests,
             'changedfiles': changedfiles,
             'clrevtomanifestrev': clrevtomanifestrev,
         }
 
         if not (generate or self._ellipses):
             # sort the nodes in storage order
             nodes = sorted(nodes, key=cl.rev)
             for node in nodes:
                 c = cl.changelogrevision(node)
                 clrevorder[node] = len(clrevorder)
                 # record the first changeset introducing this manifest version
                 manifests.setdefault(c.manifest, node)
                 # Record a complete list of potentially-changed files in
                 # this manifest.
                 changedfiles.update(c.files)
 
             return state, ()
 
         # Callback for the changelog, used to collect changed files and
         # manifest nodes.
         # Returns the linkrev node (identity in the changelog case).
         def lookupcl(x):
             c = cl.changelogrevision(x)
             clrevorder[x] = len(clrevorder)
 
             if self._ellipses:
                 # Only update manifests if x is going to be sent. Otherwise we
                 # end up with bogus linkrevs specified for manifests and
                 # we skip some manifest nodes that we should otherwise
                 # have sent.
                 if (x in self._fullclnodes
                     or cl.rev(x) in self._precomputedellipsis):
 
                     manifestnode = c.manifest
                     # Record the first changeset introducing this manifest
                     # version.
                     manifests.setdefault(manifestnode, x)
                     # Set this narrow-specific dict so we have the lowest
                     # manifest revnum to look up for this cl revnum. (Part of
                     # mapping changelog ellipsis parents to manifest ellipsis
                     # parents)
                     clrevtomanifestrev.setdefault(
                         cl.rev(x), mfl.rev(manifestnode))
                 # We can't trust the changed files list in the changeset if the
                 # client requested a shallow clone.
                 if self._isshallow:
                     changedfiles.update(mfl[c.manifest].read().keys())
                 else:
                     changedfiles.update(c.files)
             else:
                 # record the first changeset introducing this manifest version
                 manifests.setdefault(c.manifest, x)
                 # Record a complete list of potentially-changed files in
                 # this manifest.
                 changedfiles.update(c.files)
 
             return x
 
         gen = deltagroup(
             self._repo, cl, nodes, True, lookupcl,
             self._forcedeltaparentprev,
             ellipses=self._ellipses,
             topic=_('changesets'),
             clrevtolocalrev={},
             fullclnodes=self._fullclnodes,
             precomputedellipsis=self._precomputedellipsis)
 
         return state, gen
 
     def generatemanifests(self, commonrevs, clrevorder, fastpathlinkrev,
                           manifests, fnodes, source, clrevtolocalrev):
         """Returns an iterator of changegroup chunks containing manifests.
 
         `source` is unused here, but is used by extensions like remotefilelog to
         change what is sent based in pulls vs pushes, etc.
         """
         repo = self._repo
         mfl = repo.manifestlog
         tmfnodes = {'': manifests}
 
         # Callback for the manifest, used to collect linkrevs for filelog
         # revisions.
         # Returns the linkrev node (collected in lookupcl).
         def makelookupmflinknode(tree, nodes):
             if fastpathlinkrev:
                 assert not tree
                 return manifests.__getitem__
 
             def lookupmflinknode(x):
                 """Callback for looking up the linknode for manifests.
 
                 Returns the linkrev node for the specified manifest.
 
                 SIDE EFFECT:
 
                 1) fclnodes gets populated with the list of relevant
                    file nodes if we're not using fastpathlinkrev
                 2) When treemanifests are in use, collects treemanifest nodes
                    to send
 
                 Note that this means manifests must be completely sent to
                 the client before you can trust the list of files and
                 treemanifests to send.
                 """
                 clnode = nodes[x]
                 mdata = mfl.get(tree, x).readfast(shallow=True)
                 for p, n, fl in mdata.iterentries():
                     if fl == 't': # subdirectory manifest
                         subtree = tree + p + '/'
                         tmfclnodes = tmfnodes.setdefault(subtree, {})
                         tmfclnode = tmfclnodes.setdefault(n, clnode)
                         if clrevorder[clnode] < clrevorder[tmfclnode]:
                             tmfclnodes[n] = clnode
                     else:
                         f = tree + p
                         fclnodes = fnodes.setdefault(f, {})
                         fclnode = fclnodes.setdefault(n, clnode)
                         if clrevorder[clnode] < clrevorder[fclnode]:
                             fclnodes[n] = clnode
                 return clnode
             return lookupmflinknode
 
         while tmfnodes:
             tree, nodes = tmfnodes.popitem()
 
             should_visit = self._matcher.visitdir(tree[:-1] or '.')
             if tree and not should_visit:
                 continue
 
             store = mfl.getstorage(tree)
 
             if not should_visit:
                 # No nodes to send because this directory is out of
                 # the client's view of the repository (probably
                 # because of narrow clones). Do this even for the root
                 # directory (tree=='')
                 prunednodes = []
             else:
                 # Avoid sending any manifest nodes we can prove the
                 # client already has by checking linkrevs. See the
                 # related comment in generatefiles().
                 prunednodes = self._prunemanifests(store, nodes, commonrevs)
 
             if tree and not prunednodes:
                 continue
 
             lookupfn = makelookupmflinknode(tree, nodes)
 
             deltas = deltagroup(
                 self._repo, store, prunednodes, False, lookupfn,
                 self._forcedeltaparentprev,
                 ellipses=self._ellipses,
                 topic=_('manifests'),
                 clrevtolocalrev=clrevtolocalrev,
                 fullclnodes=self._fullclnodes,
                 precomputedellipsis=self._precomputedellipsis)
 
             if not self._oldmatcher.visitdir(store.tree[:-1] or '.'):
                 yield tree, deltas
             else:
                 # 'deltas' is a generator and we need to consume it even if
                 # we are not going to send it because a side-effect is that
                 # it updates tmdnodes (via lookupfn)
                 for d in deltas:
                     pass
                 if not tree:
                     yield tree, []
 
     def _prunemanifests(self, store, nodes, commonrevs):
         # This is split out as a separate method to allow filtering
         # commonrevs in extension code.
         #
         # TODO(augie): this shouldn't be required, instead we should
         # make filtering of revisions to send delegated to the store
         # layer.
         frev, flr = store.rev, store.linkrev
         return [n for n in nodes if flr(frev(n)) not in commonrevs]
 
     # The 'source' parameter is useful for extensions
     def generatefiles(self, changedfiles, commonrevs, source,
                       mfdicts, fastpathlinkrev, fnodes, clrevs):
         changedfiles = [f for f in changedfiles
                         if self._matcher(f) and not self._oldmatcher(f)]
 
         if not fastpathlinkrev:
             def normallinknodes(unused, fname):
                 return fnodes.get(fname, {})
         else:
             cln = self._repo.changelog.node
 
             def normallinknodes(store, fname):
                 flinkrev = store.linkrev
                 fnode = store.node
                 revs = ((r, flinkrev(r)) for r in store)
                 return dict((fnode(r), cln(lr))
                             for r, lr in revs if lr in clrevs)
 
         clrevtolocalrev = {}
 
         if self._isshallow:
             # In a shallow clone, the linknodes callback needs to also include
             # those file nodes that are in the manifests we sent but weren't
             # introduced by those manifests.
             commonctxs = [self._repo[c] for c in commonrevs]
             clrev = self._repo.changelog.rev
 
             def linknodes(flog, fname):
                 for c in commonctxs:
                     try:
                         fnode = c.filenode(fname)
                         clrevtolocalrev[c.rev()] = flog.rev(fnode)
                     except error.ManifestLookupError:
                         pass
                 links = normallinknodes(flog, fname)
                 if len(links) != len(mfdicts):
                     for mf, lr in mfdicts:
                         fnode = mf.get(fname, None)
                         if fnode in links:
                             links[fnode] = min(links[fnode], lr, key=clrev)
                         elif fnode:
                             links[fnode] = lr
                 return links
         else:
             linknodes = normallinknodes
 
         repo = self._repo
         progress = repo.ui.makeprogress(_('files'), unit=_('files'),
                                         total=len(changedfiles))
         for i, fname in enumerate(sorted(changedfiles)):
             filerevlog = repo.file(fname)
             if not filerevlog:
                 raise error.Abort(_("empty or missing file data for %s") %
                                   fname)
 
             clrevtolocalrev.clear()
 
             linkrevnodes = linknodes(filerevlog, fname)
             # Lookup for filenodes, we collected the linkrev nodes above in the
             # fastpath case and with lookupmf in the slowpath case.
             def lookupfilelog(x):
                 return linkrevnodes[x]
 
             frev, flr = filerevlog.rev, filerevlog.linkrev
             # Skip sending any filenode we know the client already
             # has. This avoids over-sending files relatively
             # inexpensively, so it's not a problem if we under-filter
             # here.
             filenodes = [n for n in linkrevnodes
                          if flr(frev(n)) not in commonrevs]
 
             if not filenodes:
                 continue
 
             progress.update(i + 1, item=fname)
 
             deltas = deltagroup(
                 self._repo, filerevlog, filenodes, False, lookupfilelog,
                 self._forcedeltaparentprev,
                 ellipses=self._ellipses,
                 clrevtolocalrev=clrevtolocalrev,
                 fullclnodes=self._fullclnodes,
                 precomputedellipsis=self._precomputedellipsis)
 
             yield fname, deltas
 
         progress.complete()
 
 def _makecg1packer(repo, oldmatcher, matcher, bundlecaps,
                    ellipses=False, shallow=False, ellipsisroots=None,
                    fullnodes=None):
     builddeltaheader = lambda d: _CHANGEGROUPV1_DELTA_HEADER.pack(
         d.node, d.p1node, d.p2node, d.linknode)
 
     return cgpacker(repo, oldmatcher, matcher, b'01',
                     builddeltaheader=builddeltaheader,
                     manifestsend=b'',
                     forcedeltaparentprev=True,
                     bundlecaps=bundlecaps,
                     ellipses=ellipses,
                     shallow=shallow,
                     ellipsisroots=ellipsisroots,
                     fullnodes=fullnodes)
 
 def _makecg2packer(repo, oldmatcher, matcher, bundlecaps,
                    ellipses=False, shallow=False, ellipsisroots=None,
                    fullnodes=None):
     builddeltaheader = lambda d: _CHANGEGROUPV2_DELTA_HEADER.pack(
         d.node, d.p1node, d.p2node, d.basenode, d.linknode)
 
     return cgpacker(repo, oldmatcher, matcher, b'02',
                     builddeltaheader=builddeltaheader,
                     manifestsend=b'',
                     bundlecaps=bundlecaps,
                     ellipses=ellipses,
                     shallow=shallow,
                     ellipsisroots=ellipsisroots,
                     fullnodes=fullnodes)
 
 def _makecg3packer(repo, oldmatcher, matcher, bundlecaps,
                    ellipses=False, shallow=False, ellipsisroots=None,
                    fullnodes=None):
     builddeltaheader = lambda d: _CHANGEGROUPV3_DELTA_HEADER.pack(
         d.node, d.p1node, d.p2node, d.basenode, d.linknode, d.flags)
 
     return cgpacker(repo, oldmatcher, matcher, b'03',
                     builddeltaheader=builddeltaheader,
                     manifestsend=closechunk(),
                     bundlecaps=bundlecaps,
                     ellipses=ellipses,
                     shallow=shallow,
                     ellipsisroots=ellipsisroots,
                     fullnodes=fullnodes)
 
 _packermap = {'01': (_makecg1packer, cg1unpacker),
              # cg2 adds support for exchanging generaldelta
              '02': (_makecg2packer, cg2unpacker),
              # cg3 adds support for exchanging revlog flags and treemanifests
              '03': (_makecg3packer, cg3unpacker),
 }
 
 def allsupportedversions(repo):
     versions = set(_packermap.keys())
     if not (repo.ui.configbool('experimental', 'changegroup3') or
             repo.ui.configbool('experimental', 'treemanifest') or
             'treemanifest' in repo.requirements):
         versions.discard('03')
     return versions
 
 # Changegroup versions that can be applied to the repo
 def supportedincomingversions(repo):
     return allsupportedversions(repo)
 
 # Changegroup versions that can be created from the repo
 def supportedoutgoingversions(repo):
     versions = allsupportedversions(repo)
     if 'treemanifest' in repo.requirements:
         # Versions 01 and 02 support only flat manifests and it's just too
         # expensive to convert between the flat manifest and tree manifest on
         # the fly. Since tree manifests are hashed differently, all of history
         # would have to be converted. Instead, we simply don't even pretend to
         # support versions 01 and 02.
         versions.discard('01')
         versions.discard('02')
     if repository.NARROW_REQUIREMENT in repo.requirements:
         # Versions 01 and 02 don't support revlog flags, and we need to
         # support that for stripping and unbundling to work.
         versions.discard('01')
         versions.discard('02')
     if LFS_REQUIREMENT in repo.requirements:
         # Versions 01 and 02 don't support revlog flags, and we need to
         # mark LFS entries with REVIDX_EXTSTORED.
         versions.discard('01')
         versions.discard('02')
 
     return versions
 
 def localversion(repo):
     # Finds the best version to use for bundles that are meant to be used
     # locally, such as those from strip and shelve, and temporary bundles.
     return max(supportedoutgoingversions(repo))
 
 def safeversion(repo):
     # Finds the smallest version that it's safe to assume clients of the repo
     # will support. For example, all hg versions that support generaldelta also
     # support changegroup 02.
     versions = supportedoutgoingversions(repo)
     if 'generaldelta' in repo.requirements:
         versions.discard('01')
     assert versions
     return min(versions)
 
 def getbundler(version, repo, bundlecaps=None, oldmatcher=None,
                matcher=None, ellipses=False, shallow=False,
                ellipsisroots=None, fullnodes=None):
     assert version in supportedoutgoingversions(repo)
 
     if matcher is None:
-        matcher = matchmod.alwaysmatcher(repo.root, '')
+        matcher = matchmod.always(repo.root, '')
     if oldmatcher is None:
-        oldmatcher = matchmod.nevermatcher(repo.root, '')
+        oldmatcher = matchmod.never(repo.root, '')
 
     if version == '01' and not matcher.always():
         raise error.ProgrammingError('version 01 changegroups do not support '
                                      'sparse file matchers')
 
     if ellipses and version in (b'01', b'02'):
         raise error.Abort(
             _('ellipsis nodes require at least cg3 on client and server, '
               'but negotiated version %s') % version)
 
     # Requested files could include files not in the local store. So
     # filter those out.
     matcher = repo.narrowmatch(matcher)
 
     fn = _packermap[version][0]
     return fn(repo, oldmatcher, matcher, bundlecaps, ellipses=ellipses,
               shallow=shallow, ellipsisroots=ellipsisroots,
               fullnodes=fullnodes)
 
 def getunbundler(version, fh, alg, extras=None):
     return _packermap[version][1](fh, alg, extras=extras)
 
 def _changegroupinfo(repo, nodes, source):
     if repo.ui.verbose or source == 'bundle':
         repo.ui.status(_("%d changesets found\n") % len(nodes))
     if repo.ui.debugflag:
         repo.ui.debug("list of changesets:\n")
         for node in nodes:
             repo.ui.debug("%s\n" % hex(node))
 
 def makechangegroup(repo, outgoing, version, source, fastpath=False,
                     bundlecaps=None):
     cgstream = makestream(repo, outgoing, version, source,
                           fastpath=fastpath, bundlecaps=bundlecaps)
     return getunbundler(version, util.chunkbuffer(cgstream), None,
                         {'clcount': len(outgoing.missing) })
 
 def makestream(repo, outgoing, version, source, fastpath=False,
                bundlecaps=None, matcher=None):
     bundler = getbundler(version, repo, bundlecaps=bundlecaps,
                          matcher=matcher)
 
     repo = repo.unfiltered()
     commonrevs = outgoing.common
     csets = outgoing.missing
     heads = outgoing.missingheads
     # We go through the fast path if we get told to, or if all (unfiltered
     # heads have been requested (since we then know there all linkrevs will
     # be pulled by the client).
     heads.sort()
     fastpathlinkrev = fastpath or (
             repo.filtername is None and heads == sorted(repo.heads()))
 
     repo.hook('preoutgoing', throw=True, source=source)
     _changegroupinfo(repo, csets, source)
     return bundler.generate(commonrevs, csets, fastpathlinkrev, source)
 
 def _addchangegroupfiles(repo, source, revmap, trp, expectedfiles, needfiles):
     revisions = 0
     files = 0
     progress = repo.ui.makeprogress(_('files'), unit=_('files'),
                                     total=expectedfiles)
     for chunkdata in iter(source.filelogheader, {}):
         files += 1
         f = chunkdata["filename"]
         repo.ui.debug("adding %s revisions\n" % f)
         progress.increment()
         fl = repo.file(f)
         o = len(fl)
         try:
             deltas = source.deltaiter()
             if not fl.addgroup(deltas, revmap, trp):
                 raise error.Abort(_("received file revlog group is empty"))
         except error.CensoredBaseError as e:
             raise error.Abort(_("received delta base is censored: %s") % e)
         revisions += len(fl) - o
         if f in needfiles:
             needs = needfiles[f]
             for new in pycompat.xrange(o, len(fl)):
                 n = fl.node(new)
                 if n in needs:
                     needs.remove(n)
                 else:
                     raise error.Abort(
                         _("received spurious file revlog entry"))
             if not needs:
                 del needfiles[f]
     progress.complete()
 
     for f, needs in needfiles.iteritems():
         fl = repo.file(f)
         for n in needs:
             try:
                 fl.rev(n)
             except error.LookupError:
                 raise error.Abort(
                     _('missing file data for %s:%s - run hg verify') %
                     (f, hex(n)))
 
     return revisions, files

mercurial/fileset.py

@@ -1,562 +1,561 b''
 # fileset.py - file set queries for mercurial
 #
 # Copyright 2010 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import errno
 import re
 
 from .i18n import _
 from . import (
     error,
     filesetlang,
     match as matchmod,
     merge,
     pycompat,
     registrar,
     scmutil,
     util,
 )
 from .utils import (
     stringutil,
 )
 
 # common weight constants
 _WEIGHT_CHECK_FILENAME = filesetlang.WEIGHT_CHECK_FILENAME
 _WEIGHT_READ_CONTENTS = filesetlang.WEIGHT_READ_CONTENTS
 _WEIGHT_STATUS = filesetlang.WEIGHT_STATUS
 _WEIGHT_STATUS_THOROUGH = filesetlang.WEIGHT_STATUS_THOROUGH
 
 # helpers for processing parsed tree
 getsymbol = filesetlang.getsymbol
 getstring = filesetlang.getstring
 _getkindpat = filesetlang.getkindpat
 getpattern = filesetlang.getpattern
 getargs = filesetlang.getargs
 
 def getmatch(mctx, x):
     if not x:
         raise error.ParseError(_("missing argument"))
     return methods[x[0]](mctx, *x[1:])
 
 def getmatchwithstatus(mctx, x, hint):
     keys = set(getstring(hint, 'status hint must be a string').split())
     return getmatch(mctx.withstatus(keys), x)
 
 def stringmatch(mctx, x):
     return mctx.matcher([x])
 
 def kindpatmatch(mctx, x, y):
     return stringmatch(mctx, _getkindpat(x, y, matchmod.allpatternkinds,
                                          _("pattern must be a string")))
 
 def patternsmatch(mctx, *xs):
     allkinds = matchmod.allpatternkinds
     patterns = [getpattern(x, allkinds, _("pattern must be a string"))
                 for x in xs]
     return mctx.matcher(patterns)
 
 def andmatch(mctx, x, y):
     xm = getmatch(mctx, x)
     ym = getmatch(mctx.narrowed(xm), y)
     return matchmod.intersectmatchers(xm, ym)
 
 def ormatch(mctx, *xs):
     ms = [getmatch(mctx, x) for x in xs]
     return matchmod.unionmatcher(ms)
 
 def notmatch(mctx, x):
     m = getmatch(mctx, x)
     return mctx.predicate(lambda f: not m(f), predrepr=('<not %r>', m))
 
 def minusmatch(mctx, x, y):
     xm = getmatch(mctx, x)
     ym = getmatch(mctx.narrowed(xm), y)
     return matchmod.differencematcher(xm, ym)
 
 def listmatch(mctx, *xs):
     raise error.ParseError(_("can't use a list in this context"),
                            hint=_('see \'hg help "filesets.x or y"\''))
 
 def func(mctx, a, b):
     funcname = getsymbol(a)
     if funcname in symbols:
         return symbols[funcname](mctx, b)
 
     keep = lambda fn: getattr(fn, '__doc__', None) is not None
 
     syms = [s for (s, fn) in symbols.items() if keep(fn)]
     raise error.UnknownIdentifier(funcname, syms)
 
 # symbols are callable like:
 #  fun(mctx, x)
 # with:
 #  mctx - current matchctx instance
 #  x - argument in tree form
 symbols = filesetlang.symbols
 
 predicate = registrar.filesetpredicate(symbols)
 
 @predicate('modified()', callstatus=True, weight=_WEIGHT_STATUS)
 def modified(mctx, x):
     """File that is modified according to :hg:`status`.
     """
     # i18n: "modified" is a keyword
     getargs(x, 0, 0, _("modified takes no arguments"))
     s = set(mctx.status().modified)
     return mctx.predicate(s.__contains__, predrepr='modified')
 
 @predicate('added()', callstatus=True, weight=_WEIGHT_STATUS)
 def added(mctx, x):
     """File that is added according to :hg:`status`.
     """
     # i18n: "added" is a keyword
     getargs(x, 0, 0, _("added takes no arguments"))
     s = set(mctx.status().added)
     return mctx.predicate(s.__contains__, predrepr='added')
 
 @predicate('removed()', callstatus=True, weight=_WEIGHT_STATUS)
 def removed(mctx, x):
     """File that is removed according to :hg:`status`.
     """
     # i18n: "removed" is a keyword
     getargs(x, 0, 0, _("removed takes no arguments"))
     s = set(mctx.status().removed)
     return mctx.predicate(s.__contains__, predrepr='removed')
 
 @predicate('deleted()', callstatus=True, weight=_WEIGHT_STATUS)
 def deleted(mctx, x):
     """Alias for ``missing()``.
     """
     # i18n: "deleted" is a keyword
     getargs(x, 0, 0, _("deleted takes no arguments"))
     s = set(mctx.status().deleted)
     return mctx.predicate(s.__contains__, predrepr='deleted')
 
 @predicate('missing()', callstatus=True, weight=_WEIGHT_STATUS)
 def missing(mctx, x):
     """File that is missing according to :hg:`status`.
     """
     # i18n: "missing" is a keyword
     getargs(x, 0, 0, _("missing takes no arguments"))
     s = set(mctx.status().deleted)
     return mctx.predicate(s.__contains__, predrepr='deleted')
 
 @predicate('unknown()', callstatus=True, weight=_WEIGHT_STATUS_THOROUGH)
 def unknown(mctx, x):
     """File that is unknown according to :hg:`status`."""
     # i18n: "unknown" is a keyword
     getargs(x, 0, 0, _("unknown takes no arguments"))
     s = set(mctx.status().unknown)
     return mctx.predicate(s.__contains__, predrepr='unknown')
 
 @predicate('ignored()', callstatus=True, weight=_WEIGHT_STATUS_THOROUGH)
 def ignored(mctx, x):
     """File that is ignored according to :hg:`status`."""
     # i18n: "ignored" is a keyword
     getargs(x, 0, 0, _("ignored takes no arguments"))
     s = set(mctx.status().ignored)
     return mctx.predicate(s.__contains__, predrepr='ignored')
 
 @predicate('clean()', callstatus=True, weight=_WEIGHT_STATUS)
 def clean(mctx, x):
     """File that is clean according to :hg:`status`.
     """
     # i18n: "clean" is a keyword
     getargs(x, 0, 0, _("clean takes no arguments"))
     s = set(mctx.status().clean)
     return mctx.predicate(s.__contains__, predrepr='clean')
 
 @predicate('tracked()')
 def tracked(mctx, x):
     """File that is under Mercurial control."""
     # i18n: "tracked" is a keyword
     getargs(x, 0, 0, _("tracked takes no arguments"))
     return mctx.predicate(mctx.ctx.__contains__, predrepr='tracked')
 
 @predicate('binary()', weight=_WEIGHT_READ_CONTENTS)
 def binary(mctx, x):
     """File that appears to be binary (contains NUL bytes).
     """
     # i18n: "binary" is a keyword
     getargs(x, 0, 0, _("binary takes no arguments"))
     return mctx.fpredicate(lambda fctx: fctx.isbinary(),
                            predrepr='binary', cache=True)
 
 @predicate('exec()')
 def exec_(mctx, x):
     """File that is marked as executable.
     """
     # i18n: "exec" is a keyword
     getargs(x, 0, 0, _("exec takes no arguments"))
     ctx = mctx.ctx
     return mctx.predicate(lambda f: ctx.flags(f) == 'x', predrepr='exec')
 
 @predicate('symlink()')
 def symlink(mctx, x):
     """File that is marked as a symlink.
     """
     # i18n: "symlink" is a keyword
     getargs(x, 0, 0, _("symlink takes no arguments"))
     ctx = mctx.ctx
     return mctx.predicate(lambda f: ctx.flags(f) == 'l', predrepr='symlink')
 
 @predicate('resolved()', weight=_WEIGHT_STATUS)
 def resolved(mctx, x):
     """File that is marked resolved according to :hg:`resolve -l`.
     """
     # i18n: "resolved" is a keyword
     getargs(x, 0, 0, _("resolved takes no arguments"))
     if mctx.ctx.rev() is not None:
         return mctx.never()
     ms = merge.mergestate.read(mctx.ctx.repo())
     return mctx.predicate(lambda f: f in ms and ms[f] == 'r',
                           predrepr='resolved')
 
 @predicate('unresolved()', weight=_WEIGHT_STATUS)
 def unresolved(mctx, x):
     """File that is marked unresolved according to :hg:`resolve -l`.
     """
     # i18n: "unresolved" is a keyword
     getargs(x, 0, 0, _("unresolved takes no arguments"))
     if mctx.ctx.rev() is not None:
         return mctx.never()
     ms = merge.mergestate.read(mctx.ctx.repo())
     return mctx.predicate(lambda f: f in ms and ms[f] == 'u',
                           predrepr='unresolved')
 
 @predicate('hgignore()', weight=_WEIGHT_STATUS)
 def hgignore(mctx, x):
     """File that matches the active .hgignore pattern.
     """
     # i18n: "hgignore" is a keyword
     getargs(x, 0, 0, _("hgignore takes no arguments"))
     return mctx.ctx.repo().dirstate._ignore
 
 @predicate('portable()', weight=_WEIGHT_CHECK_FILENAME)
 def portable(mctx, x):
     """File that has a portable name. (This doesn't include filenames with case
     collisions.)
     """
     # i18n: "portable" is a keyword
     getargs(x, 0, 0, _("portable takes no arguments"))
     return mctx.predicate(lambda f: util.checkwinfilename(f) is None,
                           predrepr='portable')
 
 @predicate('grep(regex)', weight=_WEIGHT_READ_CONTENTS)
 def grep(mctx, x):
     """File contains the given regular expression.
     """
     try:
         # i18n: "grep" is a keyword
         r = re.compile(getstring(x, _("grep requires a pattern")))
     except re.error as e:
         raise error.ParseError(_('invalid match pattern: %s') %
                                stringutil.forcebytestr(e))
     return mctx.fpredicate(lambda fctx: r.search(fctx.data()),
                            predrepr=('grep(%r)', r.pattern), cache=True)
 
 def _sizetomax(s):
     try:
         s = s.strip().lower()
         for k, v in util._sizeunits:
             if s.endswith(k):
                 # max(4k) = 5k - 1, max(4.5k) = 4.6k - 1
                 n = s[:-len(k)]
                 inc = 1.0
                 if "." in n:
                     inc /= 10 ** len(n.split(".")[1])
                 return int((float(n) + inc) * v) - 1
         # no extension, this is a precise value
         return int(s)
     except ValueError:
         raise error.ParseError(_("couldn't parse size: %s") % s)
 
 def sizematcher(expr):
     """Return a function(size) -> bool from the ``size()`` expression"""
     expr = expr.strip()
     if '-' in expr: # do we have a range?
         a, b = expr.split('-', 1)
         a = util.sizetoint(a)
         b = util.sizetoint(b)
         return lambda x: x >= a and x <= b
     elif expr.startswith("<="):
         a = util.sizetoint(expr[2:])
         return lambda x: x <= a
     elif expr.startswith("<"):
         a = util.sizetoint(expr[1:])
         return lambda x: x < a
     elif expr.startswith(">="):
         a = util.sizetoint(expr[2:])
         return lambda x: x >= a
     elif expr.startswith(">"):
         a = util.sizetoint(expr[1:])
         return lambda x: x > a
     else:
         a = util.sizetoint(expr)
         b = _sizetomax(expr)
         return lambda x: x >= a and x <= b
 
 @predicate('size(expression)', weight=_WEIGHT_STATUS)
 def size(mctx, x):
     """File size matches the given expression. Examples:
 
     - size('1k') - files from 1024 to 2047 bytes
     - size('< 20k') - files less than 20480 bytes
     - size('>= .5MB') - files at least 524288 bytes
     - size('4k - 1MB') - files from 4096 bytes to 1048576 bytes
     """
     # i18n: "size" is a keyword
     expr = getstring(x, _("size requires an expression"))
     m = sizematcher(expr)
     return mctx.fpredicate(lambda fctx: m(fctx.size()),
                            predrepr=('size(%r)', expr), cache=True)
 
 @predicate('encoding(name)', weight=_WEIGHT_READ_CONTENTS)
 def encoding(mctx, x):
     """File can be successfully decoded with the given character
     encoding. May not be useful for encodings other than ASCII and
     UTF-8.
     """
 
     # i18n: "encoding" is a keyword
     enc = getstring(x, _("encoding requires an encoding name"))
 
     def encp(fctx):
         d = fctx.data()
         try:
             d.decode(pycompat.sysstr(enc))
             return True
         except LookupError:
             raise error.Abort(_("unknown encoding '%s'") % enc)
         except UnicodeDecodeError:
             return False
 
     return mctx.fpredicate(encp, predrepr=('encoding(%r)', enc), cache=True)
 
 @predicate('eol(style)', weight=_WEIGHT_READ_CONTENTS)
 def eol(mctx, x):
     """File contains newlines of the given style (dos, unix, mac). Binary
     files are excluded, files with mixed line endings match multiple
     styles.
     """
 
     # i18n: "eol" is a keyword
     enc = getstring(x, _("eol requires a style name"))
 
     def eolp(fctx):
         if fctx.isbinary():
             return False
         d = fctx.data()
         if (enc == 'dos' or enc == 'win') and '\r\n' in d:
             return True
         elif enc == 'unix' and re.search('(?<!\r)\n', d):
             return True
         elif enc == 'mac' and re.search('\r(?!\n)', d):
             return True
         return False
     return mctx.fpredicate(eolp, predrepr=('eol(%r)', enc), cache=True)
 
 @predicate('copied()')
 def copied(mctx, x):
     """File that is recorded as being copied.
     """
     # i18n: "copied" is a keyword
     getargs(x, 0, 0, _("copied takes no arguments"))
     def copiedp(fctx):
         p = fctx.parents()
         return p and p[0].path() != fctx.path()
     return mctx.fpredicate(copiedp, predrepr='copied', cache=True)
 
 @predicate('revs(revs, pattern)', weight=_WEIGHT_STATUS)
 def revs(mctx, x):
     """Evaluate set in the specified revisions. If the revset match multiple
     revs, this will return file matching pattern in any of the revision.
     """
     # i18n: "revs" is a keyword
     r, x = getargs(x, 2, 2, _("revs takes two arguments"))
     # i18n: "revs" is a keyword
     revspec = getstring(r, _("first argument to revs must be a revision"))
     repo = mctx.ctx.repo()
     revs = scmutil.revrange(repo, [revspec])
 
     matchers = []
     for r in revs:
         ctx = repo[r]
         mc = mctx.switch(ctx.p1(), ctx)
         matchers.append(getmatch(mc, x))
     if not matchers:
         return mctx.never()
     if len(matchers) == 1:
         return matchers[0]
     return matchmod.unionmatcher(matchers)
 
 @predicate('status(base, rev, pattern)', weight=_WEIGHT_STATUS)
 def status(mctx, x):
     """Evaluate predicate using status change between ``base`` and
     ``rev``. Examples:
 
     - ``status(3, 7, added())`` - matches files added from "3" to "7"
     """
     repo = mctx.ctx.repo()
     # i18n: "status" is a keyword
     b, r, x = getargs(x, 3, 3, _("status takes three arguments"))
     # i18n: "status" is a keyword
     baseerr = _("first argument to status must be a revision")
     baserevspec = getstring(b, baseerr)
     if not baserevspec:
         raise error.ParseError(baseerr)
     reverr = _("second argument to status must be a revision")
     revspec = getstring(r, reverr)
     if not revspec:
         raise error.ParseError(reverr)
     basectx, ctx = scmutil.revpair(repo, [baserevspec, revspec])
     mc = mctx.switch(basectx, ctx)
     return getmatch(mc, x)
 
 @predicate('subrepo([pattern])')
 def subrepo(mctx, x):
     """Subrepositories whose paths match the given pattern.
     """
     # i18n: "subrepo" is a keyword
     getargs(x, 0, 1, _("subrepo takes at most one argument"))
     ctx = mctx.ctx
     sstate = ctx.substate
     if x:
         pat = getpattern(x, matchmod.allpatternkinds,
                          # i18n: "subrepo" is a keyword
                          _("subrepo requires a pattern or no arguments"))
         fast = not matchmod.patkind(pat)
         if fast:
             def m(s):
                 return (s == pat)
         else:
             m = matchmod.match(ctx.repo().root, '', [pat], ctx=ctx)
         return mctx.predicate(lambda f: f in sstate and m(f),
                               predrepr=('subrepo(%r)', pat))
     else:
         return mctx.predicate(sstate.__contains__, predrepr='subrepo')
 
 methods = {
     'withstatus': getmatchwithstatus,
     'string': stringmatch,
     'symbol': stringmatch,
     'kindpat': kindpatmatch,
     'patterns': patternsmatch,
     'and': andmatch,
     'or': ormatch,
     'minus': minusmatch,
     'list': listmatch,
     'not': notmatch,
     'func': func,
 }
 
 class matchctx(object):
     def __init__(self, basectx, ctx, badfn=None):
         self._basectx = basectx
         self.ctx = ctx
         self._badfn = badfn
         self._match = None
         self._status = None
 
     def narrowed(self, match):
         """Create matchctx for a sub-tree narrowed by the given matcher"""
         mctx = matchctx(self._basectx, self.ctx, self._badfn)
         mctx._match = match
         # leave wider status which we don't have to care
         mctx._status = self._status
         return mctx
 
     def switch(self, basectx, ctx):
         mctx = matchctx(basectx, ctx, self._badfn)
         mctx._match = self._match
         return mctx
 
     def withstatus(self, keys):
         """Create matchctx which has precomputed status specified by the keys"""
         mctx = matchctx(self._basectx, self.ctx, self._badfn)
         mctx._match = self._match
         mctx._buildstatus(keys)
         return mctx
 
     def _buildstatus(self, keys):
         self._status = self._basectx.status(self.ctx, self._match,
                                             listignored='ignored' in keys,
                                             listclean='clean' in keys,
                                             listunknown='unknown' in keys)
 
     def status(self):
         return self._status
 
     def matcher(self, patterns):
         return self.ctx.match(patterns, badfn=self._badfn)
 
     def predicate(self, predfn, predrepr=None, cache=False):
         """Create a matcher to select files by predfn(filename)"""
         if cache:
             predfn = util.cachefunc(predfn)
         repo = self.ctx.repo()
         return matchmod.predicatematcher(repo.root, repo.getcwd(), predfn,
                                          predrepr=predrepr, badfn=self._badfn)
 
     def fpredicate(self, predfn, predrepr=None, cache=False):
         """Create a matcher to select files by predfn(fctx) at the current
         revision
 
         Missing files are ignored.
         """
         ctx = self.ctx
         if ctx.rev() is None:
             def fctxpredfn(f):
                 try:
                     fctx = ctx[f]
                 except error.LookupError:
                     return False
                 try:
                     fctx.audit()
                 except error.Abort:
                     return False
                 try:
                     return predfn(fctx)
                 except (IOError, OSError) as e:
                     # open()-ing a directory fails with EACCES on Windows
                     if e.errno in (errno.ENOENT, errno.EACCES, errno.ENOTDIR,
                                    errno.EISDIR):
                         return False
                     raise
         else:
             def fctxpredfn(f):
                 try:
                     fctx = ctx[f]
                 except error.LookupError:
                     return False
                 return predfn(fctx)
         return self.predicate(fctxpredfn, predrepr=predrepr, cache=cache)
 
     def never(self):
         """Create a matcher to select nothing"""
         repo = self.ctx.repo()
-        return matchmod.nevermatcher(repo.root, repo.getcwd(),
-                                     badfn=self._badfn)
+        return matchmod.never(repo.root, repo.getcwd(), badfn=self._badfn)
 
 def match(ctx, expr, badfn=None):
     """Create a matcher for a single fileset expression"""
     tree = filesetlang.parse(expr)
     tree = filesetlang.analyze(tree)
     tree = filesetlang.optimize(tree)
     mctx = matchctx(ctx.p1(), ctx, badfn=badfn)
     return getmatch(mctx, tree)
 
 
 def loadpredicate(ui, extname, registrarobj):
     """Load fileset predicates from specified registrarobj
     """
     for name, func in registrarobj._table.iteritems():
         symbols[name] = func
 
 # tell hggettext to extract docstrings from these functions:
 i18nfunctions = symbols.values()

mercurial/subrepo.py

@@ -1,1840 +1,1840 b''
 # subrepo.py - sub-repository classes and factory
 #
 # Copyright 2009-2010 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import copy
 import errno
 import hashlib
 import os
 import re
 import stat
 import subprocess
 import sys
 import tarfile
 import xml.dom.minidom
 
 from .i18n import _
 from . import (
     cmdutil,
     encoding,
     error,
     exchange,
     logcmdutil,
     match as matchmod,
     node,
     pathutil,
     phases,
     pycompat,
     scmutil,
     subrepoutil,
     util,
     vfs as vfsmod,
 )
 from .utils import (
     dateutil,
     procutil,
     stringutil,
 )
 
 hg = None
 reporelpath = subrepoutil.reporelpath
 subrelpath = subrepoutil.subrelpath
 _abssource = subrepoutil._abssource
 propertycache = util.propertycache
 
 def _expandedabspath(path):
     '''
     get a path or url and if it is a path expand it and return an absolute path
     '''
     expandedpath = util.urllocalpath(util.expandpath(path))
     u = util.url(expandedpath)
     if not u.scheme:
         path = util.normpath(os.path.abspath(u.path))
     return path
 
 def _getstorehashcachename(remotepath):
     '''get a unique filename for the store hash cache of a remote repository'''
     return node.hex(hashlib.sha1(_expandedabspath(remotepath)).digest())[0:12]
 
 class SubrepoAbort(error.Abort):
     """Exception class used to avoid handling a subrepo error more than once"""
     def __init__(self, *args, **kw):
         self.subrepo = kw.pop(r'subrepo', None)
         self.cause = kw.pop(r'cause', None)
         error.Abort.__init__(self, *args, **kw)
 
 def annotatesubrepoerror(func):
     def decoratedmethod(self, *args, **kargs):
         try:
             res = func(self, *args, **kargs)
         except SubrepoAbort as ex:
             # This exception has already been handled
             raise ex
         except error.Abort as ex:
             subrepo = subrelpath(self)
             errormsg = (stringutil.forcebytestr(ex) + ' '
                         + _('(in subrepository "%s")') % subrepo)
             # avoid handling this exception by raising a SubrepoAbort exception
             raise SubrepoAbort(errormsg, hint=ex.hint, subrepo=subrepo,
                                cause=sys.exc_info())
         return res
     return decoratedmethod
 
 def _updateprompt(ui, sub, dirty, local, remote):
     if dirty:
         msg = (_(' subrepository sources for %s differ\n'
                  'use (l)ocal source (%s) or (r)emote source (%s)?'
                  '$$ &Local $$ &Remote')
                % (subrelpath(sub), local, remote))
     else:
         msg = (_(' subrepository sources for %s differ (in checked out '
                  'version)\n'
                  'use (l)ocal source (%s) or (r)emote source (%s)?'
                  '$$ &Local $$ &Remote')
                % (subrelpath(sub), local, remote))
     return ui.promptchoice(msg, 0)
 
 def _sanitize(ui, vfs, ignore):
     for dirname, dirs, names in vfs.walk():
         for i, d in enumerate(dirs):
             if d.lower() == ignore:
                 del dirs[i]
                 break
         if vfs.basename(dirname).lower() != '.hg':
             continue
         for f in names:
             if f.lower() == 'hgrc':
                 ui.warn(_("warning: removing potentially hostile 'hgrc' "
                           "in '%s'\n") % vfs.join(dirname))
                 vfs.unlink(vfs.reljoin(dirname, f))
 
 def _auditsubrepopath(repo, path):
     # sanity check for potentially unsafe paths such as '~' and '$FOO'
     if path.startswith('~') or '$' in path or util.expandpath(path) != path:
         raise error.Abort(_('subrepo path contains illegal component: %s')
                           % path)
     # auditor doesn't check if the path itself is a symlink
     pathutil.pathauditor(repo.root)(path)
     if repo.wvfs.islink(path):
         raise error.Abort(_("subrepo '%s' traverses symbolic link") % path)
 
 SUBREPO_ALLOWED_DEFAULTS = {
     'hg': True,
     'git': False,
     'svn': False,
 }
 
 def _checktype(ui, kind):
     # subrepos.allowed is a master kill switch. If disabled, subrepos are
     # disabled period.
     if not ui.configbool('subrepos', 'allowed', True):
         raise error.Abort(_('subrepos not enabled'),
                           hint=_("see 'hg help config.subrepos' for details"))
 
     default = SUBREPO_ALLOWED_DEFAULTS.get(kind, False)
     if not ui.configbool('subrepos', '%s:allowed' % kind, default):
         raise error.Abort(_('%s subrepos not allowed') % kind,
                           hint=_("see 'hg help config.subrepos' for details"))
 
     if kind not in types:
         raise error.Abort(_('unknown subrepo type %s') % kind)
 
 def subrepo(ctx, path, allowwdir=False, allowcreate=True):
     """return instance of the right subrepo class for subrepo in path"""
     # subrepo inherently violates our import layering rules
     # because it wants to make repo objects from deep inside the stack
     # so we manually delay the circular imports to not break
     # scripts that don't use our demand-loading
     global hg
     from . import hg as h
     hg = h
 
     repo = ctx.repo()
     _auditsubrepopath(repo, path)
     state = ctx.substate[path]
     _checktype(repo.ui, state[2])
     if allowwdir:
         state = (state[0], ctx.subrev(path), state[2])
     return types[state[2]](ctx, path, state[:2], allowcreate)
 
 def nullsubrepo(ctx, path, pctx):
     """return an empty subrepo in pctx for the extant subrepo in ctx"""
     # subrepo inherently violates our import layering rules
     # because it wants to make repo objects from deep inside the stack
     # so we manually delay the circular imports to not break
     # scripts that don't use our demand-loading
     global hg
     from . import hg as h
     hg = h
 
     repo = ctx.repo()
     _auditsubrepopath(repo, path)
     state = ctx.substate[path]
     _checktype(repo.ui, state[2])
     subrev = ''
     if state[2] == 'hg':
         subrev = "0" * 40
     return types[state[2]](pctx, path, (state[0], subrev), True)
 
 # subrepo classes need to implement the following abstract class:
 
 class abstractsubrepo(object):
 
     def __init__(self, ctx, path):
         """Initialize abstractsubrepo part
 
         ``ctx`` is the context referring this subrepository in the
         parent repository.
 
         ``path`` is the path to this subrepository as seen from
         innermost repository.
         """
         self.ui = ctx.repo().ui
         self._ctx = ctx
         self._path = path
 
     def addwebdirpath(self, serverpath, webconf):
         """Add the hgwebdir entries for this subrepo, and any of its subrepos.
 
         ``serverpath`` is the path component of the URL for this repo.
 
         ``webconf`` is the dictionary of hgwebdir entries.
         """
         pass
 
     def storeclean(self, path):
         """
         returns true if the repository has not changed since it was last
         cloned from or pushed to a given repository.
         """
         return False
 
     def dirty(self, ignoreupdate=False, missing=False):
         """returns true if the dirstate of the subrepo is dirty or does not
         match current stored state. If ignoreupdate is true, only check
         whether the subrepo has uncommitted changes in its dirstate.  If missing
         is true, check for deleted files.
         """
         raise NotImplementedError
 
     def dirtyreason(self, ignoreupdate=False, missing=False):
         """return reason string if it is ``dirty()``
 
         Returned string should have enough information for the message
         of exception.
 
         This returns None, otherwise.
         """
         if self.dirty(ignoreupdate=ignoreupdate, missing=missing):
             return _('uncommitted changes in subrepository "%s"'
                      ) % subrelpath(self)
 
     def bailifchanged(self, ignoreupdate=False, hint=None):
         """raise Abort if subrepository is ``dirty()``
         """
         dirtyreason = self.dirtyreason(ignoreupdate=ignoreupdate,
                                        missing=True)
         if dirtyreason:
             raise error.Abort(dirtyreason, hint=hint)
 
     def basestate(self):
         """current working directory base state, disregarding .hgsubstate
         state and working directory modifications"""
         raise NotImplementedError
 
     def checknested(self, path):
         """check if path is a subrepository within this repository"""
         return False
 
     def commit(self, text, user, date):
         """commit the current changes to the subrepo with the given
         log message. Use given user and date if possible. Return the
         new state of the subrepo.
         """
         raise NotImplementedError
 
     def phase(self, state):
         """returns phase of specified state in the subrepository.
         """
         return phases.public
 
     def remove(self):
         """remove the subrepo
 
         (should verify the dirstate is not dirty first)
         """
         raise NotImplementedError
 
     def get(self, state, overwrite=False):
         """run whatever commands are needed to put the subrepo into
         this state
         """
         raise NotImplementedError
 
     def merge(self, state):
         """merge currently-saved state with the new state."""
         raise NotImplementedError
 
     def push(self, opts):
         """perform whatever action is analogous to 'hg push'
 
         This may be a no-op on some systems.
         """
         raise NotImplementedError
 
     def add(self, ui, match, prefix, uipathfn, explicitonly, **opts):
         return []
 
     def addremove(self, matcher, prefix, uipathfn, opts):
         self.ui.warn("%s: %s" % (prefix, _("addremove is not supported")))
         return 1
 
     def cat(self, match, fm, fntemplate, prefix, **opts):
         return 1
 
     def status(self, rev2, **opts):
         return scmutil.status([], [], [], [], [], [], [])
 
     def diff(self, ui, diffopts, node2, match, prefix, **opts):
         pass
 
     def outgoing(self, ui, dest, opts):
         return 1
 
     def incoming(self, ui, source, opts):
         return 1
 
     def files(self):
         """return filename iterator"""
         raise NotImplementedError
 
     def filedata(self, name, decode):
         """return file data, optionally passed through repo decoders"""
         raise NotImplementedError
 
     def fileflags(self, name):
         """return file flags"""
         return ''
 
     def matchfileset(self, expr, badfn=None):
         """Resolve the fileset expression for this repo"""
-        return matchmod.nevermatcher(self.wvfs.base, '', badfn=badfn)
+        return matchmod.never(self.wvfs.base, '', badfn=badfn)
 
     def printfiles(self, ui, m, fm, fmt, subrepos):
         """handle the files command for this subrepo"""
         return 1
 
     def archive(self, archiver, prefix, match=None, decode=True):
         if match is not None:
             files = [f for f in self.files() if match(f)]
         else:
             files = self.files()
         total = len(files)
         relpath = subrelpath(self)
         progress = self.ui.makeprogress(_('archiving (%s)') % relpath,
                                         unit=_('files'), total=total)
         progress.update(0)
         for name in files:
             flags = self.fileflags(name)
             mode = 'x' in flags and 0o755 or 0o644
             symlink = 'l' in flags
             archiver.addfile(prefix + name, mode, symlink,
                              self.filedata(name, decode))
             progress.increment()
         progress.complete()
         return total
 
     def walk(self, match):
         '''
         walk recursively through the directory tree, finding all files
         matched by the match function
         '''
 
     def forget(self, match, prefix, uipathfn, dryrun, interactive):
         return ([], [])
 
     def removefiles(self, matcher, prefix, uipathfn, after, force, subrepos,
                     dryrun, warnings):
         """remove the matched files from the subrepository and the filesystem,
         possibly by force and/or after the file has been removed from the
         filesystem.  Return 0 on success, 1 on any warning.
         """
         warnings.append(_("warning: removefiles not implemented (%s)")
                         % self._path)
         return 1
 
     def revert(self, substate, *pats, **opts):
         self.ui.warn(_('%s: reverting %s subrepos is unsupported\n') \
             % (substate[0], substate[2]))
         return []
 
     def shortid(self, revid):
         return revid
 
     def unshare(self):
         '''
         convert this repository from shared to normal storage.
         '''
 
     def verify(self):
         '''verify the integrity of the repository.  Return 0 on success or
         warning, 1 on any error.
         '''
         return 0
 
     @propertycache
     def wvfs(self):
         """return vfs to access the working directory of this subrepository
         """
         return vfsmod.vfs(self._ctx.repo().wvfs.join(self._path))
 
     @propertycache
     def _relpath(self):
         """return path to this subrepository as seen from outermost repository
         """
         return self.wvfs.reljoin(reporelpath(self._ctx.repo()), self._path)
 
 class hgsubrepo(abstractsubrepo):
     def __init__(self, ctx, path, state, allowcreate):
         super(hgsubrepo, self).__init__(ctx, path)
         self._state = state
         r = ctx.repo()
         root = r.wjoin(path)
         create = allowcreate and not r.wvfs.exists('%s/.hg' % path)
         # repository constructor does expand variables in path, which is
         # unsafe since subrepo path might come from untrusted source.
         if os.path.realpath(util.expandpath(root)) != root:
             raise error.Abort(_('subrepo path contains illegal component: %s')
                               % path)
         self._repo = hg.repository(r.baseui, root, create=create)
         if self._repo.root != root:
             raise error.ProgrammingError('failed to reject unsafe subrepo '
                                          'path: %s (expanded to %s)'
                                          % (root, self._repo.root))
 
         # Propagate the parent's --hidden option
         if r is r.unfiltered():
             self._repo = self._repo.unfiltered()
 
         self.ui = self._repo.ui
         for s, k in [('ui', 'commitsubrepos')]:
             v = r.ui.config(s, k)
             if v:
                 self.ui.setconfig(s, k, v, 'subrepo')
         # internal config: ui._usedassubrepo
         self.ui.setconfig('ui', '_usedassubrepo', 'True', 'subrepo')
         self._initrepo(r, state[0], create)
 
     @annotatesubrepoerror
     def addwebdirpath(self, serverpath, webconf):
         cmdutil.addwebdirpath(self._repo, subrelpath(self), webconf)
 
     def storeclean(self, path):
         with self._repo.lock():
             return self._storeclean(path)
 
     def _storeclean(self, path):
         clean = True
         itercache = self._calcstorehash(path)
         for filehash in self._readstorehashcache(path):
             if filehash != next(itercache, None):
                 clean = False
                 break
         if clean:
             # if not empty:
             # the cached and current pull states have a different size
             clean = next(itercache, None) is None
         return clean
 
     def _calcstorehash(self, remotepath):
         '''calculate a unique "store hash"
 
         This method is used to to detect when there are changes that may
         require a push to a given remote path.'''
         # sort the files that will be hashed in increasing (likely) file size
         filelist = ('bookmarks', 'store/phaseroots', 'store/00changelog.i')
         yield '# %s\n' % _expandedabspath(remotepath)
         vfs = self._repo.vfs
         for relname in filelist:
             filehash = node.hex(hashlib.sha1(vfs.tryread(relname)).digest())
             yield '%s = %s\n' % (relname, filehash)
 
     @propertycache
     def _cachestorehashvfs(self):
         return vfsmod.vfs(self._repo.vfs.join('cache/storehash'))
 
     def _readstorehashcache(self, remotepath):
         '''read the store hash cache for a given remote repository'''
         cachefile = _getstorehashcachename(remotepath)
         return self._cachestorehashvfs.tryreadlines(cachefile, 'r')
 
     def _cachestorehash(self, remotepath):
         '''cache the current store hash
 
         Each remote repo requires its own store hash cache, because a subrepo
         store may be "clean" versus a given remote repo, but not versus another
         '''
         cachefile = _getstorehashcachename(remotepath)
         with self._repo.lock():
             storehash = list(self._calcstorehash(remotepath))
             vfs = self._cachestorehashvfs
             vfs.writelines(cachefile, storehash, mode='wb', notindexed=True)
 
     def _getctx(self):
         '''fetch the context for this subrepo revision, possibly a workingctx
         '''
         if self._ctx.rev() is None:
             return self._repo[None]  # workingctx if parent is workingctx
         else:
             rev = self._state[1]
             return self._repo[rev]
 
     @annotatesubrepoerror
     def _initrepo(self, parentrepo, source, create):
         self._repo._subparent = parentrepo
         self._repo._subsource = source
 
         if create:
             lines = ['[paths]\n']
 
             def addpathconfig(key, value):
                 if value:
                     lines.append('%s = %s\n' % (key, value))
                     self.ui.setconfig('paths', key, value, 'subrepo')
 
             defpath = _abssource(self._repo, abort=False)
             defpushpath = _abssource(self._repo, True, abort=False)
             addpathconfig('default', defpath)
             if defpath != defpushpath:
                 addpathconfig('default-push', defpushpath)
 
             self._repo.vfs.write('hgrc', util.tonativeeol(''.join(lines)))
 
     @annotatesubrepoerror
     def add(self, ui, match, prefix, uipathfn, explicitonly, **opts):
         return cmdutil.add(ui, self._repo, match, prefix, uipathfn,
                            explicitonly, **opts)
 
     @annotatesubrepoerror
     def addremove(self, m, prefix, uipathfn, opts):
         # In the same way as sub directories are processed, once in a subrepo,
         # always entry any of its subrepos.  Don't corrupt the options that will
         # be used to process sibling subrepos however.
         opts = copy.copy(opts)
         opts['subrepos'] = True
         return scmutil.addremove(self._repo, m, prefix, uipathfn, opts)
 
     @annotatesubrepoerror
     def cat(self, match, fm, fntemplate, prefix, **opts):
         rev = self._state[1]
         ctx = self._repo[rev]
         return cmdutil.cat(self.ui, self._repo, ctx, match, fm, fntemplate,
                            prefix, **opts)
 
     @annotatesubrepoerror
     def status(self, rev2, **opts):
         try:
             rev1 = self._state[1]
             ctx1 = self._repo[rev1]
             ctx2 = self._repo[rev2]
             return self._repo.status(ctx1, ctx2, **opts)
         except error.RepoLookupError as inst:
             self.ui.warn(_('warning: error "%s" in subrepository "%s"\n')
                          % (inst, subrelpath(self)))
             return scmutil.status([], [], [], [], [], [], [])
 
     @annotatesubrepoerror
     def diff(self, ui, diffopts, node2, match, prefix, **opts):
         try:
             node1 = node.bin(self._state[1])
             # We currently expect node2 to come from substate and be
             # in hex format
             if node2 is not None:
                 node2 = node.bin(node2)
             logcmdutil.diffordiffstat(ui, self._repo, diffopts, node1, node2,
                                       match, prefix=prefix, listsubrepos=True,
                                       **opts)
         except error.RepoLookupError as inst:
             self.ui.warn(_('warning: error "%s" in subrepository "%s"\n')
                           % (inst, subrelpath(self)))
 
     @annotatesubrepoerror
     def archive(self, archiver, prefix, match=None, decode=True):
         self._get(self._state + ('hg',))
         files = self.files()
         if match:
             files = [f for f in files if match(f)]
         rev = self._state[1]
         ctx = self._repo[rev]
         scmutil.prefetchfiles(self._repo, [ctx.rev()],
                               scmutil.matchfiles(self._repo, files))
         total = abstractsubrepo.archive(self, archiver, prefix, match)
         for subpath in ctx.substate:
             s = subrepo(ctx, subpath, True)
             submatch = matchmod.subdirmatcher(subpath, match)
             subprefix = prefix + subpath + '/'
             total += s.archive(archiver, subprefix, submatch,
                                decode)
         return total
 
     @annotatesubrepoerror
     def dirty(self, ignoreupdate=False, missing=False):
         r = self._state[1]
         if r == '' and not ignoreupdate: # no state recorded
             return True
         w = self._repo[None]
         if r != w.p1().hex() and not ignoreupdate:
             # different version checked out
             return True
         return w.dirty(missing=missing) # working directory changed
 
     def basestate(self):
         return self._repo['.'].hex()
 
     def checknested(self, path):
         return self._repo._checknested(self._repo.wjoin(path))
 
     @annotatesubrepoerror
     def commit(self, text, user, date):
         # don't bother committing in the subrepo if it's only been
         # updated
         if not self.dirty(True):
             return self._repo['.'].hex()
         self.ui.debug("committing subrepo %s\n" % subrelpath(self))
         n = self._repo.commit(text, user, date)
         if not n:
             return self._repo['.'].hex() # different version checked out
         return node.hex(n)
 
     @annotatesubrepoerror
     def phase(self, state):
         return self._repo[state or '.'].phase()
 
     @annotatesubrepoerror
     def remove(self):
         # we can't fully delete the repository as it may contain
         # local-only history
         self.ui.note(_('removing subrepo %s\n') % subrelpath(self))
         hg.clean(self._repo, node.nullid, False)
 
     def _get(self, state):
         source, revision, kind = state
         parentrepo = self._repo._subparent
 
         if revision in self._repo.unfiltered():
             # Allow shared subrepos tracked at null to setup the sharedpath
             if len(self._repo) != 0 or not parentrepo.shared():
                 return True
         self._repo._subsource = source
         srcurl = _abssource(self._repo)
 
         # Defer creating the peer until after the status message is logged, in
         # case there are network problems.
         getpeer = lambda: hg.peer(self._repo, {}, srcurl)
 
         if len(self._repo) == 0:
             # use self._repo.vfs instead of self.wvfs to remove .hg only
             self._repo.vfs.rmtree()
 
             # A remote subrepo could be shared if there is a local copy
             # relative to the parent's share source.  But clone pooling doesn't
             # assemble the repos in a tree, so that can't be consistently done.
             # A simpler option is for the user to configure clone pooling, and
             # work with that.
             if parentrepo.shared() and hg.islocal(srcurl):
                 self.ui.status(_('sharing subrepo %s from %s\n')
                                % (subrelpath(self), srcurl))
                 shared = hg.share(self._repo._subparent.baseui,
                                   getpeer(), self._repo.root,
                                   update=False, bookmarks=False)
                 self._repo = shared.local()
             else:
                 # TODO: find a common place for this and this code in the
                 # share.py wrap of the clone command.
                 if parentrepo.shared():
                     pool = self.ui.config('share', 'pool')
                     if pool:
                         pool = util.expandpath(pool)
 
                     shareopts = {
                         'pool': pool,
                         'mode': self.ui.config('share', 'poolnaming'),
                     }
                 else:
                     shareopts = {}
 
                 self.ui.status(_('cloning subrepo %s from %s\n')
                                % (subrelpath(self), util.hidepassword(srcurl)))
                 other, cloned = hg.clone(self._repo._subparent.baseui, {},
                                          getpeer(), self._repo.root,
                                          update=False, shareopts=shareopts)
                 self._repo = cloned.local()
             self._initrepo(parentrepo, source, create=True)
             self._cachestorehash(srcurl)
         else:
             self.ui.status(_('pulling subrepo %s from %s\n')
                            % (subrelpath(self), util.hidepassword(srcurl)))
             cleansub = self.storeclean(srcurl)
             exchange.pull(self._repo, getpeer())
             if cleansub:
                 # keep the repo clean after pull
                 self._cachestorehash(srcurl)
         return False
 
     @annotatesubrepoerror
     def get(self, state, overwrite=False):
         inrepo = self._get(state)
         source, revision, kind = state
         repo = self._repo
         repo.ui.debug("getting subrepo %s\n" % self._path)
         if inrepo:
             urepo = repo.unfiltered()
             ctx = urepo[revision]
             if ctx.hidden():
                 urepo.ui.warn(
                     _('revision %s in subrepository "%s" is hidden\n') \
                     % (revision[0:12], self._path))
                 repo = urepo
         hg.updaterepo(repo, revision, overwrite)
 
     @annotatesubrepoerror
     def merge(self, state):
         self._get(state)
         cur = self._repo['.']
         dst = self._repo[state[1]]
         anc = dst.ancestor(cur)
 
         def mergefunc():
             if anc == cur and dst.branch() == cur.branch():
                 self.ui.debug('updating subrepository "%s"\n'
                               % subrelpath(self))
                 hg.update(self._repo, state[1])
             elif anc == dst:
                 self.ui.debug('skipping subrepository "%s"\n'
                               % subrelpath(self))
             else:
                 self.ui.debug('merging subrepository "%s"\n' % subrelpath(self))
                 hg.merge(self._repo, state[1], remind=False)
 
         wctx = self._repo[None]
         if self.dirty():
             if anc != dst:
                 if _updateprompt(self.ui, self, wctx.dirty(), cur, dst):
                     mergefunc()
             else:
                 mergefunc()
         else:
             mergefunc()
 
     @annotatesubrepoerror
     def push(self, opts):
         force = opts.get('force')
         newbranch = opts.get('new_branch')
         ssh = opts.get('ssh')
 
         # push subrepos depth-first for coherent ordering
         c = self._repo['.']
         subs = c.substate # only repos that are committed
         for s in sorted(subs):
             if c.sub(s).push(opts) == 0:
                 return False
 
         dsturl = _abssource(self._repo, True)
         if not force:
             if self.storeclean(dsturl):
                 self.ui.status(
                     _('no changes made to subrepo %s since last push to %s\n')
                     % (subrelpath(self), util.hidepassword(dsturl)))
                 return None
         self.ui.status(_('pushing subrepo %s to %s\n') %
             (subrelpath(self), util.hidepassword(dsturl)))
         other = hg.peer(self._repo, {'ssh': ssh}, dsturl)
         res = exchange.push(self._repo, other, force, newbranch=newbranch)
 
         # the repo is now clean
         self._cachestorehash(dsturl)
         return res.cgresult
 
     @annotatesubrepoerror
     def outgoing(self, ui, dest, opts):
         if 'rev' in opts or 'branch' in opts:
             opts = copy.copy(opts)
             opts.pop('rev', None)
             opts.pop('branch', None)
         return hg.outgoing(ui, self._repo, _abssource(self._repo, True), opts)
 
     @annotatesubrepoerror
     def incoming(self, ui, source, opts):
         if 'rev' in opts or 'branch' in opts:
             opts = copy.copy(opts)
             opts.pop('rev', None)
             opts.pop('branch', None)
         return hg.incoming(ui, self._repo, _abssource(self._repo, False), opts)
 
     @annotatesubrepoerror
     def files(self):
         rev = self._state[1]
         ctx = self._repo[rev]
         return ctx.manifest().keys()
 
     def filedata(self, name, decode):
         rev = self._state[1]
         data = self._repo[rev][name].data()
         if decode:
             data = self._repo.wwritedata(name, data)
         return data
 
     def fileflags(self, name):
         rev = self._state[1]
         ctx = self._repo[rev]
         return ctx.flags(name)
 
     @annotatesubrepoerror
     def printfiles(self, ui, m, fm, fmt, subrepos):
         # If the parent context is a workingctx, use the workingctx here for
         # consistency.
         if self._ctx.rev() is None:
             ctx = self._repo[None]
         else:
             rev = self._state[1]
             ctx = self._repo[rev]
         return cmdutil.files(ui, ctx, m, fm, fmt, subrepos)
 
     @annotatesubrepoerror
     def matchfileset(self, expr, badfn=None):
         repo = self._repo
         if self._ctx.rev() is None:
             ctx = repo[None]
         else:
             rev = self._state[1]
             ctx = repo[rev]
 
         matchers = [ctx.matchfileset(expr, badfn=badfn)]
 
         for subpath in ctx.substate:
             sub = ctx.sub(subpath)
 
             try:
                 sm = sub.matchfileset(expr, badfn=badfn)
                 pm = matchmod.prefixdirmatcher(repo.root, repo.getcwd(),
                                                subpath, sm, badfn=badfn)
                 matchers.append(pm)
             except error.LookupError:
                 self.ui.status(_("skipping missing subrepository: %s\n")
                                % self.wvfs.reljoin(reporelpath(self), subpath))
         if len(matchers) == 1:
             return matchers[0]
         return matchmod.unionmatcher(matchers)
 
     def walk(self, match):
         ctx = self._repo[None]
         return ctx.walk(match)
 
     @annotatesubrepoerror
     def forget(self, match, prefix, uipathfn, dryrun, interactive):
         return cmdutil.forget(self.ui, self._repo, match, prefix, uipathfn,
                               True, dryrun=dryrun, interactive=interactive)
 
     @annotatesubrepoerror
     def removefiles(self, matcher, prefix, uipathfn, after, force, subrepos,
                     dryrun, warnings):
         return cmdutil.remove(self.ui, self._repo, matcher, prefix, uipathfn,
                               after, force, subrepos, dryrun)
 
     @annotatesubrepoerror
     def revert(self, substate, *pats, **opts):
         # reverting a subrepo is a 2 step process:
         # 1. if the no_backup is not set, revert all modified
         #    files inside the subrepo
         # 2. update the subrepo to the revision specified in
         #    the corresponding substate dictionary
         self.ui.status(_('reverting subrepo %s\n') % substate[0])
         if not opts.get(r'no_backup'):
             # Revert all files on the subrepo, creating backups
             # Note that this will not recursively revert subrepos
             # We could do it if there was a set:subrepos() predicate
             opts = opts.copy()
             opts[r'date'] = None
             opts[r'rev'] = substate[1]
 
             self.filerevert(*pats, **opts)
 
         # Update the repo to the revision specified in the given substate
         if not opts.get(r'dry_run'):
             self.get(substate, overwrite=True)
 
     def filerevert(self, *pats, **opts):
         ctx = self._repo[opts[r'rev']]
         parents = self._repo.dirstate.parents()
         if opts.get(r'all'):
             pats = ['set:modified()']
         else:
             pats = []
         cmdutil.revert(self.ui, self._repo, ctx, parents, *pats, **opts)
 
     def shortid(self, revid):
         return revid[:12]
 
     @annotatesubrepoerror
     def unshare(self):
         # subrepo inherently violates our import layering rules
         # because it wants to make repo objects from deep inside the stack
         # so we manually delay the circular imports to not break
         # scripts that don't use our demand-loading
         global hg
         from . import hg as h
         hg = h
 
         # Nothing prevents a user from sharing in a repo, and then making that a
         # subrepo.  Alternately, the previous unshare attempt may have failed
         # part way through.  So recurse whether or not this layer is shared.
         if self._repo.shared():
             self.ui.status(_("unsharing subrepo '%s'\n") % self._relpath)
 
         hg.unshare(self.ui, self._repo)
 
     def verify(self):
         try:
             rev = self._state[1]
             ctx = self._repo.unfiltered()[rev]
             if ctx.hidden():
                 # Since hidden revisions aren't pushed/pulled, it seems worth an
                 # explicit warning.
                 ui = self._repo.ui
                 ui.warn(_("subrepo '%s' is hidden in revision %s\n") %
                         (self._relpath, node.short(self._ctx.node())))
             return 0
         except error.RepoLookupError:
             # A missing subrepo revision may be a case of needing to pull it, so
             # don't treat this as an error.
             self._repo.ui.warn(_("subrepo '%s' not found in revision %s\n") %
                                (self._relpath, node.short(self._ctx.node())))
             return 0
 
     @propertycache
     def wvfs(self):
         """return own wvfs for efficiency and consistency
         """
         return self._repo.wvfs
 
     @propertycache
     def _relpath(self):
         """return path to this subrepository as seen from outermost repository
         """
         # Keep consistent dir separators by avoiding vfs.join(self._path)
         return reporelpath(self._repo)
 
 class svnsubrepo(abstractsubrepo):
     def __init__(self, ctx, path, state, allowcreate):
         super(svnsubrepo, self).__init__(ctx, path)
         self._state = state
         self._exe = procutil.findexe('svn')
         if not self._exe:
             raise error.Abort(_("'svn' executable not found for subrepo '%s'")
                              % self._path)
 
     def _svncommand(self, commands, filename='', failok=False):
         cmd = [self._exe]
         extrakw = {}
         if not self.ui.interactive():
             # Making stdin be a pipe should prevent svn from behaving
             # interactively even if we can't pass --non-interactive.
             extrakw[r'stdin'] = subprocess.PIPE
             # Starting in svn 1.5 --non-interactive is a global flag
             # instead of being per-command, but we need to support 1.4 so
             # we have to be intelligent about what commands take
             # --non-interactive.
             if commands[0] in ('update', 'checkout', 'commit'):
                 cmd.append('--non-interactive')
         cmd.extend(commands)
         if filename is not None:
             path = self.wvfs.reljoin(self._ctx.repo().origroot,
                                      self._path, filename)
             cmd.append(path)
         env = dict(encoding.environ)
         # Avoid localized output, preserve current locale for everything else.
         lc_all = env.get('LC_ALL')
         if lc_all:
             env['LANG'] = lc_all
             del env['LC_ALL']
         env['LC_MESSAGES'] = 'C'
         p = subprocess.Popen(pycompat.rapply(procutil.tonativestr, cmd),
                              bufsize=-1, close_fds=procutil.closefds,
                              stdout=subprocess.PIPE, stderr=subprocess.PIPE,
                              env=procutil.tonativeenv(env), **extrakw)
         stdout, stderr = map(util.fromnativeeol, p.communicate())
         stderr = stderr.strip()
         if not failok:
             if p.returncode:
                 raise error.Abort(stderr or 'exited with code %d'
                                   % p.returncode)
             if stderr:
                 self.ui.warn(stderr + '\n')
         return stdout, stderr
 
     @propertycache
     def _svnversion(self):
         output, err = self._svncommand(['--version', '--quiet'], filename=None)
         m = re.search(br'^(\d+)\.(\d+)', output)
         if not m:
             raise error.Abort(_('cannot retrieve svn tool version'))
         return (int(m.group(1)), int(m.group(2)))
 
     def _svnmissing(self):
         return not self.wvfs.exists('.svn')
 
     def _wcrevs(self):
         # Get the working directory revision as well as the last
         # commit revision so we can compare the subrepo state with
         # both. We used to store the working directory one.
         output, err = self._svncommand(['info', '--xml'])
         doc = xml.dom.minidom.parseString(output)
         entries = doc.getElementsByTagName(r'entry')
         lastrev, rev = '0', '0'
         if entries:
             rev = pycompat.bytestr(entries[0].getAttribute(r'revision')) or '0'
             commits = entries[0].getElementsByTagName(r'commit')
             if commits:
                 lastrev = pycompat.bytestr(
                     commits[0].getAttribute(r'revision')) or '0'
         return (lastrev, rev)
 
     def _wcrev(self):
         return self._wcrevs()[0]
 
     def _wcchanged(self):
         """Return (changes, extchanges, missing) where changes is True
         if the working directory was changed, extchanges is
         True if any of these changes concern an external entry and missing
         is True if any change is a missing entry.
         """
         output, err = self._svncommand(['status', '--xml'])
         externals, changes, missing = [], [], []
         doc = xml.dom.minidom.parseString(output)
         for e in doc.getElementsByTagName(r'entry'):
             s = e.getElementsByTagName(r'wc-status')
             if not s:
                 continue
             item = s[0].getAttribute(r'item')
             props = s[0].getAttribute(r'props')
             path = e.getAttribute(r'path').encode('utf8')
             if item == r'external':
                 externals.append(path)
             elif item == r'missing':
                 missing.append(path)
             if (item not in (r'', r'normal', r'unversioned', r'external')
                 or props not in (r'', r'none', r'normal')):
                 changes.append(path)
         for path in changes:
             for ext in externals:
                 if path == ext or path.startswith(ext + pycompat.ossep):
                     return True, True, bool(missing)
         return bool(changes), False, bool(missing)
 
     @annotatesubrepoerror
     def dirty(self, ignoreupdate=False, missing=False):
         if self._svnmissing():
             return self._state[1] != ''
         wcchanged = self._wcchanged()
         changed = wcchanged[0] or (missing and wcchanged[2])
         if not changed:
             if self._state[1] in self._wcrevs() or ignoreupdate:
                 return False
         return True
 
     def basestate(self):
         lastrev, rev = self._wcrevs()
         if lastrev != rev:
             # Last committed rev is not the same than rev. We would
             # like to take lastrev but we do not know if the subrepo
             # URL exists at lastrev.  Test it and fallback to rev it
             # is not there.
             try:
                 self._svncommand(['list', '%s@%s' % (self._state[0], lastrev)])
                 return lastrev
             except error.Abort:
                 pass
         return rev
 
     @annotatesubrepoerror
     def commit(self, text, user, date):
         # user and date are out of our hands since svn is centralized
         changed, extchanged, missing = self._wcchanged()
         if not changed:
             return self.basestate()
         if extchanged:
             # Do not try to commit externals
             raise error.Abort(_('cannot commit svn externals'))
         if missing:
             # svn can commit with missing entries but aborting like hg
             # seems a better approach.
             raise error.Abort(_('cannot commit missing svn entries'))
         commitinfo, err = self._svncommand(['commit', '-m', text])
         self.ui.status(commitinfo)
         newrev = re.search('Committed revision ([0-9]+).', commitinfo)
         if not newrev:
             if not commitinfo.strip():
                 # Sometimes, our definition of "changed" differs from
                 # svn one. For instance, svn ignores missing files
                 # when committing. If there are only missing files, no
                 # commit is made, no output and no error code.
                 raise error.Abort(_('failed to commit svn changes'))
             raise error.Abort(commitinfo.splitlines()[-1])
         newrev = newrev.groups()[0]
         self.ui.status(self._svncommand(['update', '-r', newrev])[0])
         return newrev
 
     @annotatesubrepoerror
     def remove(self):
         if self.dirty():
             self.ui.warn(_('not removing repo %s because '
                            'it has changes.\n') % self._path)
             return
         self.ui.note(_('removing subrepo %s\n') % self._path)
 
         self.wvfs.rmtree(forcibly=True)
         try:
             pwvfs = self._ctx.repo().wvfs
             pwvfs.removedirs(pwvfs.dirname(self._path))
         except OSError:
             pass
 
     @annotatesubrepoerror
     def get(self, state, overwrite=False):
         if overwrite:
             self._svncommand(['revert', '--recursive'])
         args = ['checkout']
         if self._svnversion >= (1, 5):
             args.append('--force')
         # The revision must be specified at the end of the URL to properly
         # update to a directory which has since been deleted and recreated.
         args.append('%s@%s' % (state[0], state[1]))
 
         # SEC: check that the ssh url is safe
         util.checksafessh(state[0])
 
         status, err = self._svncommand(args, failok=True)
         _sanitize(self.ui, self.wvfs, '.svn')
         if not re.search('Checked out revision [0-9]+.', status):
             if ('is already a working copy for a different URL' in err
                 and (self._wcchanged()[:2] == (False, False))):
                 # obstructed but clean working copy, so just blow it away.
                 self.remove()
                 self.get(state, overwrite=False)
                 return
             raise error.Abort((status or err).splitlines()[-1])
         self.ui.status(status)
 
     @annotatesubrepoerror
     def merge(self, state):
         old = self._state[1]
         new = state[1]
         wcrev = self._wcrev()
         if new != wcrev:
             dirty = old == wcrev or self._wcchanged()[0]
             if _updateprompt(self.ui, self, dirty, wcrev, new):
                 self.get(state, False)
 
     def push(self, opts):
         # push is a no-op for SVN
         return True
 
     @annotatesubrepoerror
     def files(self):
         output = self._svncommand(['list', '--recursive', '--xml'])[0]
         doc = xml.dom.minidom.parseString(output)
         paths = []
         for e in doc.getElementsByTagName(r'entry'):
             kind = pycompat.bytestr(e.getAttribute(r'kind'))
             if kind != 'file':
                 continue
             name = r''.join(c.data for c
                             in e.getElementsByTagName(r'name')[0].childNodes
                             if c.nodeType == c.TEXT_NODE)
             paths.append(name.encode('utf8'))
         return paths
 
     def filedata(self, name, decode):
         return self._svncommand(['cat'], name)[0]
 
 
 class gitsubrepo(abstractsubrepo):
     def __init__(self, ctx, path, state, allowcreate):
         super(gitsubrepo, self).__init__(ctx, path)
         self._state = state
         self._abspath = ctx.repo().wjoin(path)
         self._subparent = ctx.repo()
         self._ensuregit()
 
     def _ensuregit(self):
         try:
             self._gitexecutable = 'git'
             out, err = self._gitnodir(['--version'])
         except OSError as e:
             genericerror = _("error executing git for subrepo '%s': %s")
             notfoundhint = _("check git is installed and in your PATH")
             if e.errno != errno.ENOENT:
                 raise error.Abort(genericerror % (
                     self._path, encoding.strtolocal(e.strerror)))
             elif pycompat.iswindows:
                 try:
                     self._gitexecutable = 'git.cmd'
                     out, err = self._gitnodir(['--version'])
                 except OSError as e2:
                     if e2.errno == errno.ENOENT:
                         raise error.Abort(_("couldn't find 'git' or 'git.cmd'"
                             " for subrepo '%s'") % self._path,
                             hint=notfoundhint)
                     else:
                         raise error.Abort(genericerror % (self._path,
                             encoding.strtolocal(e2.strerror)))
             else:
                 raise error.Abort(_("couldn't find git for subrepo '%s'")
                     % self._path, hint=notfoundhint)
         versionstatus = self._checkversion(out)
         if versionstatus == 'unknown':
             self.ui.warn(_('cannot retrieve git version\n'))
         elif versionstatus == 'abort':
             raise error.Abort(_('git subrepo requires at least 1.6.0 or later'))
         elif versionstatus == 'warning':
             self.ui.warn(_('git subrepo requires at least 1.6.0 or later\n'))
 
     @staticmethod
     def _gitversion(out):
         m = re.search(br'^git version (\d+)\.(\d+)\.(\d+)', out)
         if m:
             return (int(m.group(1)), int(m.group(2)), int(m.group(3)))
 
         m = re.search(br'^git version (\d+)\.(\d+)', out)
         if m:
             return (int(m.group(1)), int(m.group(2)), 0)
 
         return -1
 
     @staticmethod
     def _checkversion(out):
         '''ensure git version is new enough
 
         >>> _checkversion = gitsubrepo._checkversion
         >>> _checkversion(b'git version 1.6.0')
         'ok'
         >>> _checkversion(b'git version 1.8.5')
         'ok'
         >>> _checkversion(b'git version 1.4.0')
         'abort'
         >>> _checkversion(b'git version 1.5.0')
         'warning'
         >>> _checkversion(b'git version 1.9-rc0')
         'ok'
         >>> _checkversion(b'git version 1.9.0.265.g81cdec2')
         'ok'
         >>> _checkversion(b'git version 1.9.0.GIT')
         'ok'
         >>> _checkversion(b'git version 12345')
         'unknown'
         >>> _checkversion(b'no')
         'unknown'
         '''
         version = gitsubrepo._gitversion(out)
         # git 1.4.0 can't work at all, but 1.5.X can in at least some cases,
         # despite the docstring comment.  For now, error on 1.4.0, warn on
         # 1.5.0 but attempt to continue.
         if version == -1:
             return 'unknown'
         if version < (1, 5, 0):
             return 'abort'
         elif version < (1, 6, 0):
             return 'warning'
         return 'ok'
 
     def _gitcommand(self, commands, env=None, stream=False):
         return self._gitdir(commands, env=env, stream=stream)[0]
 
     def _gitdir(self, commands, env=None, stream=False):
         return self._gitnodir(commands, env=env, stream=stream,
                               cwd=self._abspath)
 
     def _gitnodir(self, commands, env=None, stream=False, cwd=None):
         """Calls the git command
 
         The methods tries to call the git command. versions prior to 1.6.0
         are not supported and very probably fail.
         """
         self.ui.debug('%s: git %s\n' % (self._relpath, ' '.join(commands)))
         if env is None:
             env = encoding.environ.copy()
         # disable localization for Git output (issue5176)
         env['LC_ALL'] = 'C'
         # fix for Git CVE-2015-7545
         if 'GIT_ALLOW_PROTOCOL' not in env:
             env['GIT_ALLOW_PROTOCOL'] = 'file:git:http:https:ssh'
         # unless ui.quiet is set, print git's stderr,
         # which is mostly progress and useful info
         errpipe = None
         if self.ui.quiet:
             errpipe = open(os.devnull, 'w')
         if self.ui._colormode and len(commands) and commands[0] == "diff":
             # insert the argument in the front,
             # the end of git diff arguments is used for paths
             commands.insert(1, '--color')
         p = subprocess.Popen(pycompat.rapply(procutil.tonativestr,
                                              [self._gitexecutable] + commands),
                              bufsize=-1,
                              cwd=pycompat.rapply(procutil.tonativestr, cwd),
                              env=procutil.tonativeenv(env),
                              close_fds=procutil.closefds,
                              stdout=subprocess.PIPE, stderr=errpipe)
         if stream:
             return p.stdout, None
 
         retdata = p.stdout.read().strip()
         # wait for the child to exit to avoid race condition.
         p.wait()
 
         if p.returncode != 0 and p.returncode != 1:
             # there are certain error codes that are ok
             command = commands[0]
             if command in ('cat-file', 'symbolic-ref'):
                 return retdata, p.returncode
             # for all others, abort
             raise error.Abort(_('git %s error %d in %s') %
                              (command, p.returncode, self._relpath))
 
         return retdata, p.returncode
 
     def _gitmissing(self):
         return not self.wvfs.exists('.git')
 
     def _gitstate(self):
         return self._gitcommand(['rev-parse', 'HEAD'])
 
     def _gitcurrentbranch(self):
         current, err = self._gitdir(['symbolic-ref', 'HEAD', '--quiet'])
         if err:
             current = None
         return current
 
     def _gitremote(self, remote):
         out = self._gitcommand(['remote', 'show', '-n', remote])
         line = out.split('\n')[1]
         i = line.index('URL: ') + len('URL: ')
         return line[i:]
 
     def _githavelocally(self, revision):
         out, code = self._gitdir(['cat-file', '-e', revision])
         return code == 0
 
     def _gitisancestor(self, r1, r2):
         base = self._gitcommand(['merge-base', r1, r2])
         return base == r1
 
     def _gitisbare(self):
         return self._gitcommand(['config', '--bool', 'core.bare']) == 'true'
 
     def _gitupdatestat(self):
         """This must be run before git diff-index.
         diff-index only looks at changes to file stat;
         this command looks at file contents and updates the stat."""
         self._gitcommand(['update-index', '-q', '--refresh'])
 
     def _gitbranchmap(self):
         '''returns 2 things:
         a map from git branch to revision
         a map from revision to branches'''
         branch2rev = {}
         rev2branch = {}
 
         out = self._gitcommand(['for-each-ref', '--format',
                                 '%(objectname) %(refname)'])
         for line in out.split('\n'):
             revision, ref = line.split(' ')
             if (not ref.startswith('refs/heads/') and
                 not ref.startswith('refs/remotes/')):
                 continue
             if ref.startswith('refs/remotes/') and ref.endswith('/HEAD'):
                 continue # ignore remote/HEAD redirects
             branch2rev[ref] = revision
             rev2branch.setdefault(revision, []).append(ref)
         return branch2rev, rev2branch
 
     def _gittracking(self, branches):
         'return map of remote branch to local tracking branch'
         # assumes no more than one local tracking branch for each remote
         tracking = {}
         for b in branches:
             if b.startswith('refs/remotes/'):
                 continue
             bname = b.split('/', 2)[2]
             remote = self._gitcommand(['config', 'branch.%s.remote' % bname])
             if remote:
                 ref = self._gitcommand(['config', 'branch.%s.merge' % bname])
                 tracking['refs/remotes/%s/%s' %
                          (remote, ref.split('/', 2)[2])] = b
         return tracking
 
     def _abssource(self, source):
         if '://' not in source:
             # recognize the scp syntax as an absolute source
             colon = source.find(':')
             if colon != -1 and '/' not in source[:colon]:
                 return source
         self._subsource = source
         return _abssource(self)
 
     def _fetch(self, source, revision):
         if self._gitmissing():
             # SEC: check for safe ssh url
             util.checksafessh(source)
 
             source = self._abssource(source)
             self.ui.status(_('cloning subrepo %s from %s\n') %
                             (self._relpath, source))
             self._gitnodir(['clone', source, self._abspath])
         if self._githavelocally(revision):
             return
         self.ui.status(_('pulling subrepo %s from %s\n') %
                         (self._relpath, self._gitremote('origin')))
         # try only origin: the originally cloned repo
         self._gitcommand(['fetch'])
         if not self._githavelocally(revision):
             raise error.Abort(_('revision %s does not exist in subrepository '
                                 '"%s"\n') % (revision, self._relpath))
 
     @annotatesubrepoerror
     def dirty(self, ignoreupdate=False, missing=False):
         if self._gitmissing():
             return self._state[1] != ''
         if self._gitisbare():
             return True
         if not ignoreupdate and self._state[1] != self._gitstate():
             # different version checked out
             return True
         # check for staged changes or modified files; ignore untracked files
         self._gitupdatestat()
         out, code = self._gitdir(['diff-index', '--quiet', 'HEAD'])
         return code == 1
 
     def basestate(self):
         return self._gitstate()
 
     @annotatesubrepoerror
     def get(self, state, overwrite=False):
         source, revision, kind = state
         if not revision:
             self.remove()
             return
         self._fetch(source, revision)
         # if the repo was set to be bare, unbare it
         if self._gitisbare():
             self._gitcommand(['config', 'core.bare', 'false'])
             if self._gitstate() == revision:
                 self._gitcommand(['reset', '--hard', 'HEAD'])
                 return
         elif self._gitstate() == revision:
             if overwrite:
                 # first reset the index to unmark new files for commit, because
                 # reset --hard will otherwise throw away files added for commit,
                 # not just unmark them.
                 self._gitcommand(['reset', 'HEAD'])
                 self._gitcommand(['reset', '--hard', 'HEAD'])
             return
         branch2rev, rev2branch = self._gitbranchmap()
 
         def checkout(args):
             cmd = ['checkout']
             if overwrite:
                 # first reset the index to unmark new files for commit, because
                 # the -f option will otherwise throw away files added for
                 # commit, not just unmark them.
                 self._gitcommand(['reset', 'HEAD'])
                 cmd.append('-f')
             self._gitcommand(cmd + args)
             _sanitize(self.ui, self.wvfs, '.git')
 
         def rawcheckout():
             # no branch to checkout, check it out with no branch
             self.ui.warn(_('checking out detached HEAD in '
                            'subrepository "%s"\n') % self._relpath)
             self.ui.warn(_('check out a git branch if you intend '
                             'to make changes\n'))
             checkout(['-q', revision])
 
         if revision not in rev2branch:
             rawcheckout()
             return
         branches = rev2branch[revision]
         firstlocalbranch = None
         for b in branches:
             if b == 'refs/heads/master':
                 # master trumps all other branches
                 checkout(['refs/heads/master'])
                 return
             if not firstlocalbranch and not b.startswith('refs/remotes/'):
                 firstlocalbranch = b
         if firstlocalbranch:
             checkout([firstlocalbranch])
             return
 
         tracking = self._gittracking(branch2rev.keys())
         # choose a remote branch already tracked if possible
         remote = branches[0]
         if remote not in tracking:
             for b in branches:
                 if b in tracking:
                     remote = b
                     break
 
         if remote not in tracking:
             # create a new local tracking branch
             local = remote.split('/', 3)[3]
             checkout(['-b', local, remote])
         elif self._gitisancestor(branch2rev[tracking[remote]], remote):
             # When updating to a tracked remote branch,
             # if the local tracking branch is downstream of it,
             # a normal `git pull` would have performed a "fast-forward merge"
             # which is equivalent to updating the local branch to the remote.
             # Since we are only looking at branching at update, we need to
             # detect this situation and perform this action lazily.
             if tracking[remote] != self._gitcurrentbranch():
                 checkout([tracking[remote]])
             self._gitcommand(['merge', '--ff', remote])
             _sanitize(self.ui, self.wvfs, '.git')
         else:
             # a real merge would be required, just checkout the revision
             rawcheckout()
 
     @annotatesubrepoerror
     def commit(self, text, user, date):
         if self._gitmissing():
             raise error.Abort(_("subrepo %s is missing") % self._relpath)
         cmd = ['commit', '-a', '-m', text]
         env = encoding.environ.copy()
         if user:
             cmd += ['--author', user]
         if date:
             # git's date parser silently ignores when seconds < 1e9
             # convert to ISO8601
             env['GIT_AUTHOR_DATE'] = dateutil.datestr(date,
                                                   '%Y-%m-%dT%H:%M:%S %1%2')
         self._gitcommand(cmd, env=env)
         # make sure commit works otherwise HEAD might not exist under certain
         # circumstances
         return self._gitstate()
 
     @annotatesubrepoerror
     def merge(self, state):
         source, revision, kind = state
         self._fetch(source, revision)
         base = self._gitcommand(['merge-base', revision, self._state[1]])
         self._gitupdatestat()
         out, code = self._gitdir(['diff-index', '--quiet', 'HEAD'])
 
         def mergefunc():
             if base == revision:
                 self.get(state) # fast forward merge
             elif base != self._state[1]:
                 self._gitcommand(['merge', '--no-commit', revision])
             _sanitize(self.ui, self.wvfs, '.git')
 
         if self.dirty():
             if self._gitstate() != revision:
                 dirty = self._gitstate() == self._state[1] or code != 0
                 if _updateprompt(self.ui, self, dirty,
                                  self._state[1][:7], revision[:7]):
                     mergefunc()
         else:
             mergefunc()
 
     @annotatesubrepoerror
     def push(self, opts):
         force = opts.get('force')
 
         if not self._state[1]:
             return True
         if self._gitmissing():
             raise error.Abort(_("subrepo %s is missing") % self._relpath)
         # if a branch in origin contains the revision, nothing to do
         branch2rev, rev2branch = self._gitbranchmap()
         if self._state[1] in rev2branch:
             for b in rev2branch[self._state[1]]:
                 if b.startswith('refs/remotes/origin/'):
                     return True
         for b, revision in branch2rev.iteritems():
             if b.startswith('refs/remotes/origin/'):
                 if self._gitisancestor(self._state[1], revision):
                     return True
         # otherwise, try to push the currently checked out branch
         cmd = ['push']
         if force:
             cmd.append('--force')
 
         current = self._gitcurrentbranch()
         if current:
             # determine if the current branch is even useful
             if not self._gitisancestor(self._state[1], current):
                 self.ui.warn(_('unrelated git branch checked out '
                                'in subrepository "%s"\n') % self._relpath)
                 return False
             self.ui.status(_('pushing branch %s of subrepository "%s"\n') %
                            (current.split('/', 2)[2], self._relpath))
             ret = self._gitdir(cmd + ['origin', current])
             return ret[1] == 0
         else:
             self.ui.warn(_('no branch checked out in subrepository "%s"\n'
                            'cannot push revision %s\n') %
                           (self._relpath, self._state[1]))
             return False
 
     @annotatesubrepoerror
     def add(self, ui, match, prefix, uipathfn, explicitonly, **opts):
         if self._gitmissing():
             return []
 
         s = self.status(None, unknown=True, clean=True)
 
         tracked = set()
         # dirstates 'amn' warn, 'r' is added again
         for l in (s.modified, s.added, s.deleted, s.clean):
             tracked.update(l)
 
         # Unknown files not of interest will be rejected by the matcher
         files = s.unknown
         files.extend(match.files())
 
         rejected = []
 
         files = [f for f in sorted(set(files)) if match(f)]
         for f in files:
             exact = match.exact(f)
             command = ["add"]
             if exact:
                 command.append("-f") #should be added, even if ignored
             if ui.verbose or not exact:
                 ui.status(_('adding %s\n') % uipathfn(f))
 
             if f in tracked:  # hg prints 'adding' even if already tracked
                 if exact:
                     rejected.append(f)
                 continue
             if not opts.get(r'dry_run'):
                 self._gitcommand(command + [f])
 
         for f in rejected:
             ui.warn(_("%s already tracked!\n") % uipathfn(f))
 
         return rejected
 
     @annotatesubrepoerror
     def remove(self):
         if self._gitmissing():
             return
         if self.dirty():
             self.ui.warn(_('not removing repo %s because '
                            'it has changes.\n') % self._relpath)
             return
         # we can't fully delete the repository as it may contain
         # local-only history
         self.ui.note(_('removing subrepo %s\n') % self._relpath)
         self._gitcommand(['config', 'core.bare', 'true'])
         for f, kind in self.wvfs.readdir():
             if f == '.git':
                 continue
             if kind == stat.S_IFDIR:
                 self.wvfs.rmtree(f)
             else:
                 self.wvfs.unlink(f)
 
     def archive(self, archiver, prefix, match=None, decode=True):
         total = 0
         source, revision = self._state
         if not revision:
             return total
         self._fetch(source, revision)
 
         # Parse git's native archive command.
         # This should be much faster than manually traversing the trees
         # and objects with many subprocess calls.
         tarstream = self._gitcommand(['archive', revision], stream=True)
         tar = tarfile.open(fileobj=tarstream, mode=r'r|')
         relpath = subrelpath(self)
         progress = self.ui.makeprogress(_('archiving (%s)') % relpath,
                                         unit=_('files'))
         progress.update(0)
         for info in tar:
             if info.isdir():
                 continue
             bname = pycompat.fsencode(info.name)
             if match and not match(bname):
                 continue
             if info.issym():
                 data = info.linkname
             else:
                 data = tar.extractfile(info).read()
             archiver.addfile(prefix + bname, info.mode, info.issym(), data)
             total += 1
             progress.increment()
         progress.complete()
         return total
 
 
     @annotatesubrepoerror
     def cat(self, match, fm, fntemplate, prefix, **opts):
         rev = self._state[1]
         if match.anypats():
             return 1 #No support for include/exclude yet
 
         if not match.files():
             return 1
 
         # TODO: add support for non-plain formatter (see cmdutil.cat())
         for f in match.files():
             output = self._gitcommand(["show", "%s:%s" % (rev, f)])
             fp = cmdutil.makefileobj(self._ctx, fntemplate,
                                      pathname=self.wvfs.reljoin(prefix, f))
             fp.write(output)
             fp.close()
         return 0
 
 
     @annotatesubrepoerror
     def status(self, rev2, **opts):
         rev1 = self._state[1]
         if self._gitmissing() or not rev1:
             # if the repo is missing, return no results
             return scmutil.status([], [], [], [], [], [], [])
         modified, added, removed = [], [], []
         self._gitupdatestat()
         if rev2:
             command = ['diff-tree', '--no-renames', '-r', rev1, rev2]
         else:
             command = ['diff-index', '--no-renames', rev1]
         out = self._gitcommand(command)
         for line in out.split('\n'):
             tab = line.find('\t')
             if tab == -1:
                 continue
             status, f = line[tab - 1:tab], line[tab + 1:]
             if status == 'M':
                 modified.append(f)
             elif status == 'A':
                 added.append(f)
             elif status == 'D':
                 removed.append(f)
 
         deleted, unknown, ignored, clean = [], [], [], []
 
         command = ['status', '--porcelain', '-z']
         if opts.get(r'unknown'):
             command += ['--untracked-files=all']
         if opts.get(r'ignored'):
             command += ['--ignored']
         out = self._gitcommand(command)
 
         changedfiles = set()
         changedfiles.update(modified)
         changedfiles.update(added)
         changedfiles.update(removed)
         for line in out.split('\0'):
             if not line:
                 continue
             st = line[0:2]
             #moves and copies show 2 files on one line
             if line.find('\0') >= 0:
                 filename1, filename2 = line[3:].split('\0')
             else:
                 filename1 = line[3:]
                 filename2 = None
 
             changedfiles.add(filename1)
             if filename2:
                 changedfiles.add(filename2)
 
             if st == '??':
                 unknown.append(filename1)
             elif st == '!!':
                 ignored.append(filename1)
 
         if opts.get(r'clean'):
             out = self._gitcommand(['ls-files'])
             for f in out.split('\n'):
                 if not f in changedfiles:
                     clean.append(f)
 
         return scmutil.status(modified, added, removed, deleted,
                               unknown, ignored, clean)
 
     @annotatesubrepoerror
     def diff(self, ui, diffopts, node2, match, prefix, **opts):
         node1 = self._state[1]
         cmd = ['diff', '--no-renames']
         if opts[r'stat']:
             cmd.append('--stat')
         else:
             # for Git, this also implies '-p'
             cmd.append('-U%d' % diffopts.context)
 
         if diffopts.noprefix:
             cmd.extend(['--src-prefix=%s/' % prefix,
                         '--dst-prefix=%s/' % prefix])
         else:
             cmd.extend(['--src-prefix=a/%s/' % prefix,
                         '--dst-prefix=b/%s/' % prefix])
 
         if diffopts.ignorews:
             cmd.append('--ignore-all-space')
         if diffopts.ignorewsamount:
             cmd.append('--ignore-space-change')
         if self._gitversion(self._gitcommand(['--version'])) >= (1, 8, 4) \
                 and diffopts.ignoreblanklines:
             cmd.append('--ignore-blank-lines')
 
         cmd.append(node1)
         if node2:
             cmd.append(node2)
 
         output = ""
         if match.always():
             output += self._gitcommand(cmd) + '\n'
         else:
             st = self.status(node2)[:3]
             files = [f for sublist in st for f in sublist]
             for f in files:
                 if match(f):
                     output += self._gitcommand(cmd + ['--', f]) + '\n'
 
         if output.strip():
             ui.write(output)
 
     @annotatesubrepoerror
     def revert(self, substate, *pats, **opts):
         self.ui.status(_('reverting subrepo %s\n') % substate[0])
         if not opts.get(r'no_backup'):
             status = self.status(None)
             names = status.modified
             for name in names:
                 # backuppath() expects a path relative to the parent repo (the
                 # repo that ui.origbackuppath is relative to)
                 parentname = os.path.join(self._path, name)
                 bakname = scmutil.backuppath(self.ui, self._subparent,
                                              parentname)
                 self.ui.note(_('saving current version of %s as %s\n') %
                         (name, os.path.relpath(bakname)))
                 util.rename(self.wvfs.join(name), bakname)
 
         if not opts.get(r'dry_run'):
             self.get(substate, overwrite=True)
         return []
 
     def shortid(self, revid):
         return revid[:7]
 
 types = {
     'hg': hgsubrepo,
     'svn': svnsubrepo,
     'git': gitsubrepo,
     }